diff --git a/.gitattributes b/.gitattributes index a6344aac8c09253b3b630fb776ae94478aa0275b..a09db2ea4d1bd1fc1c09f6fca45e8e4054953535 100644 --- a/.gitattributes +++ b/.gitattributes @@ -33,3 +33,5 @@ saved_model/**/* filter=lfs diff=lfs merge=lfs -text *.zip filter=lfs diff=lfs merge=lfs -text *.zst filter=lfs diff=lfs merge=lfs -text *tfevents* filter=lfs diff=lfs merge=lfs -text +tokenizer.json filter=lfs diff=lfs merge=lfs -text +model.safetensors.index.json filter=lfs diff=lfs merge=lfs -text diff --git a/.quant_summary.txt b/.quant_summary.txt new file mode 100644 index 0000000000000000000000000000000000000000..4eab1219ef8c763266b71b6781275061b6968caf --- /dev/null +++ b/.quant_summary.txt @@ -0,0 +1,40972 @@ +model.embed_tokens.weight_quantizer TensorQuantizer(disabled) +model.embed_tokens.input_quantizer HardDisabledTensorQuantizer(disabled) +model.embed_tokens.output_quantizer TensorQuantizer(disabled) +model.layers.0.self_attn.q_a_proj.input_quantizer TensorQuantizer(disabled) +model.layers.0.self_attn.q_a_proj.output_quantizer TensorQuantizer(disabled) +model.layers.0.self_attn.q_a_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.0.self_attn.q_b_proj.input_quantizer TensorQuantizer(disabled) +model.layers.0.self_attn.q_b_proj.output_quantizer TensorQuantizer(disabled) +model.layers.0.self_attn.q_b_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.0.self_attn.kv_a_proj_with_mqa.input_quantizer TensorQuantizer(disabled) +model.layers.0.self_attn.kv_a_proj_with_mqa.output_quantizer TensorQuantizer(disabled) +model.layers.0.self_attn.kv_a_proj_with_mqa.weight_quantizer TensorQuantizer(disabled) +model.layers.0.self_attn.kv_b_proj.input_quantizer TensorQuantizer(disabled) +model.layers.0.self_attn.kv_b_proj.output_quantizer TensorQuantizer(disabled) +model.layers.0.self_attn.kv_b_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.0.self_attn.o_proj.input_quantizer TensorQuantizer(disabled) +model.layers.0.self_attn.o_proj.output_quantizer TensorQuantizer(disabled) +model.layers.0.self_attn.o_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.0.self_attn.indexer.wq_b.input_quantizer TensorQuantizer(disabled) +model.layers.0.self_attn.indexer.wq_b.output_quantizer TensorQuantizer(disabled) +model.layers.0.self_attn.indexer.wq_b.weight_quantizer TensorQuantizer(disabled) +model.layers.0.self_attn.indexer.wk.input_quantizer TensorQuantizer(disabled) +model.layers.0.self_attn.indexer.wk.output_quantizer TensorQuantizer(disabled) +model.layers.0.self_attn.indexer.wk.weight_quantizer TensorQuantizer(disabled) +model.layers.0.self_attn.indexer.k_norm.input_quantizer TensorQuantizer(disabled) +model.layers.0.self_attn.indexer.k_norm.output_quantizer TensorQuantizer(disabled) +model.layers.0.self_attn.indexer.weights_proj.input_quantizer TensorQuantizer(disabled) +model.layers.0.self_attn.indexer.weights_proj.output_quantizer TensorQuantizer(disabled) +model.layers.0.self_attn.indexer.weights_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.0.self_attn.q_bmm_quantizer TensorQuantizer(disabled) +model.layers.0.self_attn.k_bmm_quantizer TensorQuantizer((4, 3) bit fake per-tensor amax=4.48e+02(const) calibrator=MaxCalibrator quant) +model.layers.0.self_attn.v_bmm_quantizer TensorQuantizer((4, 3) bit fake per-tensor amax=4.48e+02(const) calibrator=MaxCalibrator quant) +model.layers.0.self_attn.p_bmm_quantizer TensorQuantizer(disabled) +model.layers.0.mlp.gate_proj.input_quantizer TensorQuantizer(disabled) +model.layers.0.mlp.gate_proj.output_quantizer TensorQuantizer(disabled) +model.layers.0.mlp.gate_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.0.mlp.up_proj.input_quantizer TensorQuantizer(disabled) +model.layers.0.mlp.up_proj.output_quantizer TensorQuantizer(disabled) +model.layers.0.mlp.up_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.0.mlp.down_proj.input_quantizer TensorQuantizer(disabled) +model.layers.0.mlp.down_proj.output_quantizer TensorQuantizer(disabled) +model.layers.0.mlp.down_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.1.self_attn.q_a_proj.input_quantizer TensorQuantizer(disabled) +model.layers.1.self_attn.q_a_proj.output_quantizer TensorQuantizer(disabled) +model.layers.1.self_attn.q_a_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.1.self_attn.q_b_proj.input_quantizer TensorQuantizer(disabled) +model.layers.1.self_attn.q_b_proj.output_quantizer TensorQuantizer(disabled) +model.layers.1.self_attn.q_b_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.1.self_attn.kv_a_proj_with_mqa.input_quantizer TensorQuantizer(disabled) +model.layers.1.self_attn.kv_a_proj_with_mqa.output_quantizer TensorQuantizer(disabled) +model.layers.1.self_attn.kv_a_proj_with_mqa.weight_quantizer TensorQuantizer(disabled) +model.layers.1.self_attn.kv_b_proj.input_quantizer TensorQuantizer(disabled) +model.layers.1.self_attn.kv_b_proj.output_quantizer TensorQuantizer(disabled) +model.layers.1.self_attn.kv_b_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.1.self_attn.o_proj.input_quantizer TensorQuantizer(disabled) +model.layers.1.self_attn.o_proj.output_quantizer TensorQuantizer(disabled) +model.layers.1.self_attn.o_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.1.self_attn.indexer.wq_b.input_quantizer TensorQuantizer(disabled) +model.layers.1.self_attn.indexer.wq_b.output_quantizer TensorQuantizer(disabled) +model.layers.1.self_attn.indexer.wq_b.weight_quantizer TensorQuantizer(disabled) +model.layers.1.self_attn.indexer.wk.input_quantizer TensorQuantizer(disabled) +model.layers.1.self_attn.indexer.wk.output_quantizer TensorQuantizer(disabled) +model.layers.1.self_attn.indexer.wk.weight_quantizer TensorQuantizer(disabled) +model.layers.1.self_attn.indexer.k_norm.input_quantizer TensorQuantizer(disabled) +model.layers.1.self_attn.indexer.k_norm.output_quantizer TensorQuantizer(disabled) +model.layers.1.self_attn.indexer.weights_proj.input_quantizer TensorQuantizer(disabled) +model.layers.1.self_attn.indexer.weights_proj.output_quantizer TensorQuantizer(disabled) +model.layers.1.self_attn.indexer.weights_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.1.self_attn.q_bmm_quantizer TensorQuantizer(disabled) +model.layers.1.self_attn.k_bmm_quantizer TensorQuantizer((4, 3) bit fake per-tensor amax=4.48e+02(const) calibrator=MaxCalibrator quant) +model.layers.1.self_attn.v_bmm_quantizer TensorQuantizer((4, 3) bit fake per-tensor amax=4.48e+02(const) calibrator=MaxCalibrator quant) +model.layers.1.self_attn.p_bmm_quantizer TensorQuantizer(disabled) +model.layers.1.mlp.gate_proj.input_quantizer TensorQuantizer(disabled) +model.layers.1.mlp.gate_proj.output_quantizer TensorQuantizer(disabled) +model.layers.1.mlp.gate_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.1.mlp.up_proj.input_quantizer TensorQuantizer(disabled) +model.layers.1.mlp.up_proj.output_quantizer TensorQuantizer(disabled) +model.layers.1.mlp.up_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.1.mlp.down_proj.input_quantizer TensorQuantizer(disabled) +model.layers.1.mlp.down_proj.output_quantizer TensorQuantizer(disabled) +model.layers.1.mlp.down_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.2.self_attn.q_a_proj.input_quantizer TensorQuantizer(disabled) +model.layers.2.self_attn.q_a_proj.output_quantizer TensorQuantizer(disabled) +model.layers.2.self_attn.q_a_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.2.self_attn.q_b_proj.input_quantizer TensorQuantizer(disabled) +model.layers.2.self_attn.q_b_proj.output_quantizer TensorQuantizer(disabled) +model.layers.2.self_attn.q_b_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.2.self_attn.kv_a_proj_with_mqa.input_quantizer TensorQuantizer(disabled) +model.layers.2.self_attn.kv_a_proj_with_mqa.output_quantizer TensorQuantizer(disabled) +model.layers.2.self_attn.kv_a_proj_with_mqa.weight_quantizer TensorQuantizer(disabled) +model.layers.2.self_attn.kv_b_proj.input_quantizer TensorQuantizer(disabled) +model.layers.2.self_attn.kv_b_proj.output_quantizer TensorQuantizer(disabled) +model.layers.2.self_attn.kv_b_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.2.self_attn.o_proj.input_quantizer TensorQuantizer(disabled) +model.layers.2.self_attn.o_proj.output_quantizer TensorQuantizer(disabled) +model.layers.2.self_attn.o_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.2.self_attn.indexer.wq_b.input_quantizer TensorQuantizer(disabled) +model.layers.2.self_attn.indexer.wq_b.output_quantizer TensorQuantizer(disabled) +model.layers.2.self_attn.indexer.wq_b.weight_quantizer TensorQuantizer(disabled) +model.layers.2.self_attn.indexer.wk.input_quantizer TensorQuantizer(disabled) +model.layers.2.self_attn.indexer.wk.output_quantizer TensorQuantizer(disabled) +model.layers.2.self_attn.indexer.wk.weight_quantizer TensorQuantizer(disabled) +model.layers.2.self_attn.indexer.k_norm.input_quantizer TensorQuantizer(disabled) +model.layers.2.self_attn.indexer.k_norm.output_quantizer TensorQuantizer(disabled) +model.layers.2.self_attn.indexer.weights_proj.input_quantizer TensorQuantizer(disabled) +model.layers.2.self_attn.indexer.weights_proj.output_quantizer TensorQuantizer(disabled) +model.layers.2.self_attn.indexer.weights_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.2.self_attn.q_bmm_quantizer TensorQuantizer(disabled) +model.layers.2.self_attn.k_bmm_quantizer TensorQuantizer((4, 3) bit fake per-tensor amax=4.48e+02(const) calibrator=MaxCalibrator quant) +model.layers.2.self_attn.v_bmm_quantizer TensorQuantizer((4, 3) bit fake per-tensor amax=4.48e+02(const) calibrator=MaxCalibrator quant) +model.layers.2.self_attn.p_bmm_quantizer TensorQuantizer(disabled) +model.layers.2.mlp.gate_proj.input_quantizer TensorQuantizer(disabled) +model.layers.2.mlp.gate_proj.output_quantizer TensorQuantizer(disabled) +model.layers.2.mlp.gate_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.2.mlp.up_proj.input_quantizer TensorQuantizer(disabled) +model.layers.2.mlp.up_proj.output_quantizer TensorQuantizer(disabled) +model.layers.2.mlp.up_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.2.mlp.down_proj.input_quantizer TensorQuantizer(disabled) +model.layers.2.mlp.down_proj.output_quantizer TensorQuantizer(disabled) +model.layers.2.mlp.down_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.3.self_attn.q_a_proj.input_quantizer TensorQuantizer(disabled) +model.layers.3.self_attn.q_a_proj.output_quantizer TensorQuantizer(disabled) +model.layers.3.self_attn.q_a_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.3.self_attn.q_b_proj.input_quantizer TensorQuantizer(disabled) +model.layers.3.self_attn.q_b_proj.output_quantizer TensorQuantizer(disabled) +model.layers.3.self_attn.q_b_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.3.self_attn.kv_a_proj_with_mqa.input_quantizer TensorQuantizer(disabled) +model.layers.3.self_attn.kv_a_proj_with_mqa.output_quantizer TensorQuantizer(disabled) +model.layers.3.self_attn.kv_a_proj_with_mqa.weight_quantizer TensorQuantizer(disabled) +model.layers.3.self_attn.kv_b_proj.input_quantizer TensorQuantizer(disabled) +model.layers.3.self_attn.kv_b_proj.output_quantizer TensorQuantizer(disabled) +model.layers.3.self_attn.kv_b_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.3.self_attn.o_proj.input_quantizer TensorQuantizer(disabled) +model.layers.3.self_attn.o_proj.output_quantizer TensorQuantizer(disabled) +model.layers.3.self_attn.o_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.3.self_attn.q_bmm_quantizer TensorQuantizer(disabled) +model.layers.3.self_attn.k_bmm_quantizer TensorQuantizer((4, 3) bit fake per-tensor amax=4.48e+02(const) calibrator=MaxCalibrator quant) +model.layers.3.self_attn.v_bmm_quantizer TensorQuantizer((4, 3) bit fake per-tensor amax=4.48e+02(const) calibrator=MaxCalibrator quant) +model.layers.3.self_attn.p_bmm_quantizer TensorQuantizer(disabled) +model.layers.3.mlp.experts.gate_up_proj_input_quantizer TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.77e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.gate_up_proj_weight_quantizers.0 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.01e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.gate_up_proj_weight_quantizers.1 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.35e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.gate_up_proj_weight_quantizers.2 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.89e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.gate_up_proj_weight_quantizers.3 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.77e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.gate_up_proj_weight_quantizers.4 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.54e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.gate_up_proj_weight_quantizers.5 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.84e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.gate_up_proj_weight_quantizers.6 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.58e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.gate_up_proj_weight_quantizers.7 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.52e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.gate_up_proj_weight_quantizers.8 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.97e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.gate_up_proj_weight_quantizers.9 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.43e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.gate_up_proj_weight_quantizers.10 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.78e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.gate_up_proj_weight_quantizers.11 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.88e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.gate_up_proj_weight_quantizers.12 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.98e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.gate_up_proj_weight_quantizers.13 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.69e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.gate_up_proj_weight_quantizers.14 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.49e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.gate_up_proj_weight_quantizers.15 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.95e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.gate_up_proj_weight_quantizers.16 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.67e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.gate_up_proj_weight_quantizers.17 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.67e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.gate_up_proj_weight_quantizers.18 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.73e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.gate_up_proj_weight_quantizers.19 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.79e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.gate_up_proj_weight_quantizers.20 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.64e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.gate_up_proj_weight_quantizers.21 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.67e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.gate_up_proj_weight_quantizers.22 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.79e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.gate_up_proj_weight_quantizers.23 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.47e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.gate_up_proj_weight_quantizers.24 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.34e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.gate_up_proj_weight_quantizers.25 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.59e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.gate_up_proj_weight_quantizers.26 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.63e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.gate_up_proj_weight_quantizers.27 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.56e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.gate_up_proj_weight_quantizers.28 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.56e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.gate_up_proj_weight_quantizers.29 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.74e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.gate_up_proj_weight_quantizers.30 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.76e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.gate_up_proj_weight_quantizers.31 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.23e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.gate_up_proj_weight_quantizers.32 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.73e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.gate_up_proj_weight_quantizers.33 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.72e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.gate_up_proj_weight_quantizers.34 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.85e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.gate_up_proj_weight_quantizers.35 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.49e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.gate_up_proj_weight_quantizers.36 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.64e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.gate_up_proj_weight_quantizers.37 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.64e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.gate_up_proj_weight_quantizers.38 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.05e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.gate_up_proj_weight_quantizers.39 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.32e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.gate_up_proj_weight_quantizers.40 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.38e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.gate_up_proj_weight_quantizers.41 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.00e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.gate_up_proj_weight_quantizers.42 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.93e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.gate_up_proj_weight_quantizers.43 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.92e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.gate_up_proj_weight_quantizers.44 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.85e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.gate_up_proj_weight_quantizers.45 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.70e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.gate_up_proj_weight_quantizers.46 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.20e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.gate_up_proj_weight_quantizers.47 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.58e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.gate_up_proj_weight_quantizers.48 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.19e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.gate_up_proj_weight_quantizers.49 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.95e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.gate_up_proj_weight_quantizers.50 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.74e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.gate_up_proj_weight_quantizers.51 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.46e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.gate_up_proj_weight_quantizers.52 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.67e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.gate_up_proj_weight_quantizers.53 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.63e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.gate_up_proj_weight_quantizers.54 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.54e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.gate_up_proj_weight_quantizers.55 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.gate_up_proj_weight_quantizers.56 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.46e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.gate_up_proj_weight_quantizers.57 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.43e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.gate_up_proj_weight_quantizers.58 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.74e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.gate_up_proj_weight_quantizers.59 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.72e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.gate_up_proj_weight_quantizers.60 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.79e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.gate_up_proj_weight_quantizers.61 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.19e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.gate_up_proj_weight_quantizers.62 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.87e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.gate_up_proj_weight_quantizers.63 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.89e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.gate_up_proj_weight_quantizers.64 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.67e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.gate_up_proj_weight_quantizers.65 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.40e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.gate_up_proj_weight_quantizers.66 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.17e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.gate_up_proj_weight_quantizers.67 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.10e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.gate_up_proj_weight_quantizers.68 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.77e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.gate_up_proj_weight_quantizers.69 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.60e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.gate_up_proj_weight_quantizers.70 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.49e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.gate_up_proj_weight_quantizers.71 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.25e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.gate_up_proj_weight_quantizers.72 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.46e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.gate_up_proj_weight_quantizers.73 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.29e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.gate_up_proj_weight_quantizers.74 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.78e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.gate_up_proj_weight_quantizers.75 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.31e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.gate_up_proj_weight_quantizers.76 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.79e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.gate_up_proj_weight_quantizers.77 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=4.71e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.gate_up_proj_weight_quantizers.78 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.85e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.gate_up_proj_weight_quantizers.79 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.50e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.gate_up_proj_weight_quantizers.80 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.72e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.gate_up_proj_weight_quantizers.81 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.55e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.gate_up_proj_weight_quantizers.82 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=5.35e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.gate_up_proj_weight_quantizers.83 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.95e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.gate_up_proj_weight_quantizers.84 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.gate_up_proj_weight_quantizers.85 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.17e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.gate_up_proj_weight_quantizers.86 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.48e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.gate_up_proj_weight_quantizers.87 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.95e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.gate_up_proj_weight_quantizers.88 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.27e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.gate_up_proj_weight_quantizers.89 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.89e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.gate_up_proj_weight_quantizers.90 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.81e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.gate_up_proj_weight_quantizers.91 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.85e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.gate_up_proj_weight_quantizers.92 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.97e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.gate_up_proj_weight_quantizers.93 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.87e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.gate_up_proj_weight_quantizers.94 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.03e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.gate_up_proj_weight_quantizers.95 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.54e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.gate_up_proj_weight_quantizers.96 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.97e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.gate_up_proj_weight_quantizers.97 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.66e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.gate_up_proj_weight_quantizers.98 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=4.16e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.gate_up_proj_weight_quantizers.99 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.66e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.gate_up_proj_weight_quantizers.100 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.61e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.gate_up_proj_weight_quantizers.101 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.29e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.gate_up_proj_weight_quantizers.102 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.36e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.gate_up_proj_weight_quantizers.103 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.69e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.gate_up_proj_weight_quantizers.104 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.08e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.gate_up_proj_weight_quantizers.105 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.15e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.gate_up_proj_weight_quantizers.106 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.48e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.gate_up_proj_weight_quantizers.107 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.47e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.gate_up_proj_weight_quantizers.108 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.53e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.gate_up_proj_weight_quantizers.109 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.79e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.gate_up_proj_weight_quantizers.110 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.98e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.gate_up_proj_weight_quantizers.111 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.40e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.gate_up_proj_weight_quantizers.112 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.46e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.gate_up_proj_weight_quantizers.113 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.19e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.gate_up_proj_weight_quantizers.114 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.97e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.gate_up_proj_weight_quantizers.115 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.31e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.gate_up_proj_weight_quantizers.116 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.24e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.gate_up_proj_weight_quantizers.117 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.58e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.gate_up_proj_weight_quantizers.118 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.65e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.gate_up_proj_weight_quantizers.119 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.46e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.gate_up_proj_weight_quantizers.120 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.51e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.gate_up_proj_weight_quantizers.121 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.51e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.gate_up_proj_weight_quantizers.122 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.68e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.gate_up_proj_weight_quantizers.123 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.86e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.gate_up_proj_weight_quantizers.124 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.70e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.gate_up_proj_weight_quantizers.125 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.90e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.gate_up_proj_weight_quantizers.126 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.82e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.gate_up_proj_weight_quantizers.127 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.66e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.gate_up_proj_weight_quantizers.128 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.61e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.gate_up_proj_weight_quantizers.129 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.51e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.gate_up_proj_weight_quantizers.130 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.01e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.gate_up_proj_weight_quantizers.131 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.75e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.gate_up_proj_weight_quantizers.132 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.74e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.gate_up_proj_weight_quantizers.133 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.00e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.gate_up_proj_weight_quantizers.134 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.67e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.gate_up_proj_weight_quantizers.135 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.15e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.gate_up_proj_weight_quantizers.136 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.73e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.gate_up_proj_weight_quantizers.137 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.79e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.gate_up_proj_weight_quantizers.138 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.69e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.gate_up_proj_weight_quantizers.139 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.79e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.gate_up_proj_weight_quantizers.140 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.46e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.gate_up_proj_weight_quantizers.141 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.40e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.gate_up_proj_weight_quantizers.142 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.93e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.gate_up_proj_weight_quantizers.143 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.23e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.gate_up_proj_weight_quantizers.144 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.80e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.gate_up_proj_weight_quantizers.145 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.56e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.gate_up_proj_weight_quantizers.146 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.gate_up_proj_weight_quantizers.147 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.75e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.gate_up_proj_weight_quantizers.148 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.09e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.gate_up_proj_weight_quantizers.149 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.46e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.gate_up_proj_weight_quantizers.150 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.74e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.gate_up_proj_weight_quantizers.151 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.gate_up_proj_weight_quantizers.152 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.12e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.gate_up_proj_weight_quantizers.153 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.28e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.gate_up_proj_weight_quantizers.154 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.60e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.gate_up_proj_weight_quantizers.155 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.10e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.gate_up_proj_weight_quantizers.156 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.07e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.gate_up_proj_weight_quantizers.157 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.60e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.gate_up_proj_weight_quantizers.158 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.46e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.gate_up_proj_weight_quantizers.159 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.35e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.gate_up_proj_weight_quantizers.160 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.44e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.gate_up_proj_weight_quantizers.161 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.62e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.gate_up_proj_weight_quantizers.162 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.73e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.gate_up_proj_weight_quantizers.163 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.49e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.gate_up_proj_weight_quantizers.164 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.gate_up_proj_weight_quantizers.165 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.62e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.gate_up_proj_weight_quantizers.166 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.42e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.gate_up_proj_weight_quantizers.167 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.59e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.gate_up_proj_weight_quantizers.168 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.96e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.gate_up_proj_weight_quantizers.169 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.47e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.gate_up_proj_weight_quantizers.170 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.46e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.gate_up_proj_weight_quantizers.171 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.85e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.gate_up_proj_weight_quantizers.172 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.59e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.gate_up_proj_weight_quantizers.173 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.88e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.gate_up_proj_weight_quantizers.174 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.53e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.gate_up_proj_weight_quantizers.175 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.60e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.gate_up_proj_weight_quantizers.176 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.94e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.gate_up_proj_weight_quantizers.177 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.84e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.gate_up_proj_weight_quantizers.178 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.83e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.gate_up_proj_weight_quantizers.179 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.53e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.gate_up_proj_weight_quantizers.180 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.67e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.gate_up_proj_weight_quantizers.181 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.73e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.gate_up_proj_weight_quantizers.182 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.34e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.gate_up_proj_weight_quantizers.183 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.66e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.gate_up_proj_weight_quantizers.184 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.03e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.gate_up_proj_weight_quantizers.185 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.96e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.gate_up_proj_weight_quantizers.186 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.78e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.gate_up_proj_weight_quantizers.187 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.95e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.gate_up_proj_weight_quantizers.188 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.65e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.gate_up_proj_weight_quantizers.189 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.97e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.gate_up_proj_weight_quantizers.190 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.60e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.gate_up_proj_weight_quantizers.191 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.52e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.gate_up_proj_weight_quantizers.192 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.54e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.gate_up_proj_weight_quantizers.193 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.81e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.gate_up_proj_weight_quantizers.194 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.72e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.gate_up_proj_weight_quantizers.195 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.72e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.gate_up_proj_weight_quantizers.196 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.81e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.gate_up_proj_weight_quantizers.197 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.67e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.gate_up_proj_weight_quantizers.198 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.68e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.gate_up_proj_weight_quantizers.199 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.44e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.gate_up_proj_weight_quantizers.200 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=4.00e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.gate_up_proj_weight_quantizers.201 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.34e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.gate_up_proj_weight_quantizers.202 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.92e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.gate_up_proj_weight_quantizers.203 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.51e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.gate_up_proj_weight_quantizers.204 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.88e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.gate_up_proj_weight_quantizers.205 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.53e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.gate_up_proj_weight_quantizers.206 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.71e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.gate_up_proj_weight_quantizers.207 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.23e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.gate_up_proj_weight_quantizers.208 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.01e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.gate_up_proj_weight_quantizers.209 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.68e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.gate_up_proj_weight_quantizers.210 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.91e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.gate_up_proj_weight_quantizers.211 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.83e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.gate_up_proj_weight_quantizers.212 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.37e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.gate_up_proj_weight_quantizers.213 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.01e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.gate_up_proj_weight_quantizers.214 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.42e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.gate_up_proj_weight_quantizers.215 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.74e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.gate_up_proj_weight_quantizers.216 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.27e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.gate_up_proj_weight_quantizers.217 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.33e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.gate_up_proj_weight_quantizers.218 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.51e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.gate_up_proj_weight_quantizers.219 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.15e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.gate_up_proj_weight_quantizers.220 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.54e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.gate_up_proj_weight_quantizers.221 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.16e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.gate_up_proj_weight_quantizers.222 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.26e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.gate_up_proj_weight_quantizers.223 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.54e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.gate_up_proj_weight_quantizers.224 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.70e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.gate_up_proj_weight_quantizers.225 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.21e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.gate_up_proj_weight_quantizers.226 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.50e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.gate_up_proj_weight_quantizers.227 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.72e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.gate_up_proj_weight_quantizers.228 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.93e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.gate_up_proj_weight_quantizers.229 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.11e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.gate_up_proj_weight_quantizers.230 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.24e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.gate_up_proj_weight_quantizers.231 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.44e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.gate_up_proj_weight_quantizers.232 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.gate_up_proj_weight_quantizers.233 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.50e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.gate_up_proj_weight_quantizers.234 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.gate_up_proj_weight_quantizers.235 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.77e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.gate_up_proj_weight_quantizers.236 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.95e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.gate_up_proj_weight_quantizers.237 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.21e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.gate_up_proj_weight_quantizers.238 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.76e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.gate_up_proj_weight_quantizers.239 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.15e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.gate_up_proj_weight_quantizers.240 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.49e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.gate_up_proj_weight_quantizers.241 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.59e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.gate_up_proj_weight_quantizers.242 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.34e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.gate_up_proj_weight_quantizers.243 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.69e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.gate_up_proj_weight_quantizers.244 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.73e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.gate_up_proj_weight_quantizers.245 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.44e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.gate_up_proj_weight_quantizers.246 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.52e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.gate_up_proj_weight_quantizers.247 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.38e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.gate_up_proj_weight_quantizers.248 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.76e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.gate_up_proj_weight_quantizers.249 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.71e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.gate_up_proj_weight_quantizers.250 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.21e+00 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.gate_up_proj_weight_quantizers.251 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.51e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.gate_up_proj_weight_quantizers.252 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.gate_up_proj_weight_quantizers.253 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.63e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.gate_up_proj_weight_quantizers.254 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.51e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.gate_up_proj_weight_quantizers.255 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.16e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.down_proj_input_quantizer TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e+00 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.down_proj_weight_quantizers.0 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.66e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.down_proj_weight_quantizers.1 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.18e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.down_proj_weight_quantizers.2 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.33e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.down_proj_weight_quantizers.3 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.45e-02 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.down_proj_weight_quantizers.4 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.down_proj_weight_quantizers.5 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.down_proj_weight_quantizers.6 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.down_proj_weight_quantizers.7 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.down_proj_weight_quantizers.8 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.73e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.down_proj_weight_quantizers.9 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.45e-02 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.down_proj_weight_quantizers.10 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.35e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.down_proj_weight_quantizers.11 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.down_proj_weight_quantizers.12 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.30e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.down_proj_weight_quantizers.13 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.down_proj_weight_quantizers.14 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.23e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.down_proj_weight_quantizers.15 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.20e-02 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.down_proj_weight_quantizers.16 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.down_proj_weight_quantizers.17 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.down_proj_weight_quantizers.18 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.05e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.down_proj_weight_quantizers.19 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.14e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.down_proj_weight_quantizers.20 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.down_proj_weight_quantizers.21 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.down_proj_weight_quantizers.22 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.14e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.down_proj_weight_quantizers.23 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.down_proj_weight_quantizers.24 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.23e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.down_proj_weight_quantizers.25 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.down_proj_weight_quantizers.26 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.50e-02 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.down_proj_weight_quantizers.27 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.down_proj_weight_quantizers.28 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.35e-02 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.down_proj_weight_quantizers.29 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.38e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.down_proj_weight_quantizers.30 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.down_proj_weight_quantizers.31 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.down_proj_weight_quantizers.32 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.54e-02 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.down_proj_weight_quantizers.33 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.40e-02 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.down_proj_weight_quantizers.34 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.down_proj_weight_quantizers.35 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.45e-02 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.down_proj_weight_quantizers.36 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.42e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.down_proj_weight_quantizers.37 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.down_proj_weight_quantizers.38 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.63e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.down_proj_weight_quantizers.39 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.14e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.down_proj_weight_quantizers.40 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.63e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.down_proj_weight_quantizers.41 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.30e-02 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.down_proj_weight_quantizers.42 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.down_proj_weight_quantizers.43 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.86e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.down_proj_weight_quantizers.44 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.51e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.down_proj_weight_quantizers.45 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.59e-02 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.down_proj_weight_quantizers.46 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.25e-02 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.down_proj_weight_quantizers.47 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.down_proj_weight_quantizers.48 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.down_proj_weight_quantizers.49 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.87e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.down_proj_weight_quantizers.50 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.down_proj_weight_quantizers.51 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.65e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.down_proj_weight_quantizers.52 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.down_proj_weight_quantizers.53 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.06e-02 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.down_proj_weight_quantizers.54 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.down_proj_weight_quantizers.55 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.down_proj_weight_quantizers.56 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.down_proj_weight_quantizers.57 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.64e-02 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.down_proj_weight_quantizers.58 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.down_proj_weight_quantizers.59 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.down_proj_weight_quantizers.60 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.35e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.down_proj_weight_quantizers.61 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.59e-02 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.down_proj_weight_quantizers.62 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.01e-02 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.down_proj_weight_quantizers.63 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.50e-02 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.down_proj_weight_quantizers.64 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.down_proj_weight_quantizers.65 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.down_proj_weight_quantizers.66 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.down_proj_weight_quantizers.67 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.71e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.down_proj_weight_quantizers.68 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.down_proj_weight_quantizers.69 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.22e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.down_proj_weight_quantizers.70 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.down_proj_weight_quantizers.71 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.22e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.down_proj_weight_quantizers.72 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.down_proj_weight_quantizers.73 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.down_proj_weight_quantizers.74 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.29e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.down_proj_weight_quantizers.75 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.down_proj_weight_quantizers.76 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.35e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.down_proj_weight_quantizers.77 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.48e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.down_proj_weight_quantizers.78 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.down_proj_weight_quantizers.79 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.down_proj_weight_quantizers.80 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.25e-02 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.down_proj_weight_quantizers.81 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.45e-02 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.down_proj_weight_quantizers.82 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.down_proj_weight_quantizers.83 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.54e-02 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.down_proj_weight_quantizers.84 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.down_proj_weight_quantizers.85 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.61e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.down_proj_weight_quantizers.86 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.64e-02 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.down_proj_weight_quantizers.87 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.83e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.down_proj_weight_quantizers.88 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.21e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.down_proj_weight_quantizers.89 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.70e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.down_proj_weight_quantizers.90 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.down_proj_weight_quantizers.91 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.34e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.down_proj_weight_quantizers.92 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.60e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.down_proj_weight_quantizers.93 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.down_proj_weight_quantizers.94 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.51e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.down_proj_weight_quantizers.95 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.44e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.down_proj_weight_quantizers.96 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.down_proj_weight_quantizers.97 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.down_proj_weight_quantizers.98 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.down_proj_weight_quantizers.99 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.down_proj_weight_quantizers.100 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.down_proj_weight_quantizers.101 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.29e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.down_proj_weight_quantizers.102 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.down_proj_weight_quantizers.103 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.40e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.down_proj_weight_quantizers.104 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.down_proj_weight_quantizers.105 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.down_proj_weight_quantizers.106 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.down_proj_weight_quantizers.107 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.down_proj_weight_quantizers.108 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.60e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.down_proj_weight_quantizers.109 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.64e-02 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.down_proj_weight_quantizers.110 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.16e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.down_proj_weight_quantizers.111 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.down_proj_weight_quantizers.112 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.50e-02 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.down_proj_weight_quantizers.113 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.down_proj_weight_quantizers.114 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.down_proj_weight_quantizers.115 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.50e-02 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.down_proj_weight_quantizers.116 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.down_proj_weight_quantizers.117 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.15e-02 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.down_proj_weight_quantizers.118 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.18e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.down_proj_weight_quantizers.119 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.45e-02 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.down_proj_weight_quantizers.120 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.down_proj_weight_quantizers.121 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.59e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.down_proj_weight_quantizers.122 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.35e-02 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.down_proj_weight_quantizers.123 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.down_proj_weight_quantizers.124 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.40e-02 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.down_proj_weight_quantizers.125 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.44e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.down_proj_weight_quantizers.126 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.19e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.down_proj_weight_quantizers.127 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.down_proj_weight_quantizers.128 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.down_proj_weight_quantizers.129 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.35e-02 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.down_proj_weight_quantizers.130 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.61e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.down_proj_weight_quantizers.131 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.47e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.down_proj_weight_quantizers.132 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.down_proj_weight_quantizers.133 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.27e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.down_proj_weight_quantizers.134 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.down_proj_weight_quantizers.135 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.31e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.down_proj_weight_quantizers.136 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.down_proj_weight_quantizers.137 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.79e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.down_proj_weight_quantizers.138 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.19e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.down_proj_weight_quantizers.139 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.down_proj_weight_quantizers.140 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.down_proj_weight_quantizers.141 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.68e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.down_proj_weight_quantizers.142 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.85e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.down_proj_weight_quantizers.143 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.03e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.down_proj_weight_quantizers.144 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.down_proj_weight_quantizers.145 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.50e-02 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.down_proj_weight_quantizers.146 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.down_proj_weight_quantizers.147 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.73e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.down_proj_weight_quantizers.148 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.85e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.down_proj_weight_quantizers.149 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.down_proj_weight_quantizers.150 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.46e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.down_proj_weight_quantizers.151 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.down_proj_weight_quantizers.152 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.15e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.down_proj_weight_quantizers.153 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.down_proj_weight_quantizers.154 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.down_proj_weight_quantizers.155 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.44e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.down_proj_weight_quantizers.156 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.40e-02 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.down_proj_weight_quantizers.157 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.50e-02 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.down_proj_weight_quantizers.158 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.down_proj_weight_quantizers.159 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.47e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.down_proj_weight_quantizers.160 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.down_proj_weight_quantizers.161 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.down_proj_weight_quantizers.162 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.down_proj_weight_quantizers.163 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.down_proj_weight_quantizers.164 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.down_proj_weight_quantizers.165 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.down_proj_weight_quantizers.166 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.35e-02 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.down_proj_weight_quantizers.167 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.down_proj_weight_quantizers.168 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.62e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.down_proj_weight_quantizers.169 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.35e-02 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.down_proj_weight_quantizers.170 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.down_proj_weight_quantizers.171 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.down_proj_weight_quantizers.172 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.down_proj_weight_quantizers.173 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.23e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.down_proj_weight_quantizers.174 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.down_proj_weight_quantizers.175 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.45e-02 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.down_proj_weight_quantizers.176 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.42e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.down_proj_weight_quantizers.177 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.down_proj_weight_quantizers.178 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.55e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.down_proj_weight_quantizers.179 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.65e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.down_proj_weight_quantizers.180 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.21e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.down_proj_weight_quantizers.181 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.down_proj_weight_quantizers.182 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.45e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.down_proj_weight_quantizers.183 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.40e-02 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.down_proj_weight_quantizers.184 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.25e-02 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.down_proj_weight_quantizers.185 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.down_proj_weight_quantizers.186 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.down_proj_weight_quantizers.187 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.24e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.down_proj_weight_quantizers.188 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.down_proj_weight_quantizers.189 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.56e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.down_proj_weight_quantizers.190 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.down_proj_weight_quantizers.191 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.down_proj_weight_quantizers.192 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.down_proj_weight_quantizers.193 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.59e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.down_proj_weight_quantizers.194 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.46e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.down_proj_weight_quantizers.195 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.down_proj_weight_quantizers.196 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.down_proj_weight_quantizers.197 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=7.91e-02 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.down_proj_weight_quantizers.198 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.down_proj_weight_quantizers.199 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.40e-02 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.down_proj_weight_quantizers.200 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.40e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.down_proj_weight_quantizers.201 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.21e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.down_proj_weight_quantizers.202 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.64e-02 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.down_proj_weight_quantizers.203 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.down_proj_weight_quantizers.204 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.64e-02 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.down_proj_weight_quantizers.205 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.35e-02 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.down_proj_weight_quantizers.206 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.down_proj_weight_quantizers.207 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.down_proj_weight_quantizers.208 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.down_proj_weight_quantizers.209 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.down_proj_weight_quantizers.210 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.30e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.down_proj_weight_quantizers.211 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.down_proj_weight_quantizers.212 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.17e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.down_proj_weight_quantizers.213 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.down_proj_weight_quantizers.214 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.down_proj_weight_quantizers.215 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.down_proj_weight_quantizers.216 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.33e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.down_proj_weight_quantizers.217 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.down_proj_weight_quantizers.218 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.35e-02 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.down_proj_weight_quantizers.219 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.77e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.down_proj_weight_quantizers.220 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.down_proj_weight_quantizers.221 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.down_proj_weight_quantizers.222 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.64e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.down_proj_weight_quantizers.223 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.down_proj_weight_quantizers.224 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.45e-02 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.down_proj_weight_quantizers.225 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.down_proj_weight_quantizers.226 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.down_proj_weight_quantizers.227 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.47e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.down_proj_weight_quantizers.228 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.71e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.down_proj_weight_quantizers.229 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.08e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.down_proj_weight_quantizers.230 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.00e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.down_proj_weight_quantizers.231 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.69e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.down_proj_weight_quantizers.232 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.down_proj_weight_quantizers.233 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.down_proj_weight_quantizers.234 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.down_proj_weight_quantizers.235 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.34e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.down_proj_weight_quantizers.236 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.84e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.down_proj_weight_quantizers.237 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.31e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.down_proj_weight_quantizers.238 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.40e-02 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.down_proj_weight_quantizers.239 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.06e-02 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.down_proj_weight_quantizers.240 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.26e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.down_proj_weight_quantizers.241 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.down_proj_weight_quantizers.242 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.54e-02 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.down_proj_weight_quantizers.243 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.down_proj_weight_quantizers.244 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.down_proj_weight_quantizers.245 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.59e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.down_proj_weight_quantizers.246 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.03e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.down_proj_weight_quantizers.247 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.29e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.down_proj_weight_quantizers.248 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.25e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.down_proj_weight_quantizers.249 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.down_proj_weight_quantizers.250 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.09e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.down_proj_weight_quantizers.251 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.down_proj_weight_quantizers.252 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.down_proj_weight_quantizers.253 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.down_proj_weight_quantizers.254 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.down_proj_weight_quantizers.255 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.3.mlp.shared_experts.gate_proj.input_quantizer TensorQuantizer(disabled) +model.layers.3.mlp.shared_experts.gate_proj.output_quantizer TensorQuantizer(disabled) +model.layers.3.mlp.shared_experts.gate_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.3.mlp.shared_experts.up_proj.input_quantizer TensorQuantizer(disabled) +model.layers.3.mlp.shared_experts.up_proj.output_quantizer TensorQuantizer(disabled) +model.layers.3.mlp.shared_experts.up_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.3.mlp.shared_experts.down_proj.input_quantizer TensorQuantizer(disabled) +model.layers.3.mlp.shared_experts.down_proj.output_quantizer TensorQuantizer(disabled) +model.layers.3.mlp.shared_experts.down_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.4.self_attn.q_a_proj.input_quantizer TensorQuantizer(disabled) +model.layers.4.self_attn.q_a_proj.output_quantizer TensorQuantizer(disabled) +model.layers.4.self_attn.q_a_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.4.self_attn.q_b_proj.input_quantizer TensorQuantizer(disabled) +model.layers.4.self_attn.q_b_proj.output_quantizer TensorQuantizer(disabled) +model.layers.4.self_attn.q_b_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.4.self_attn.kv_a_proj_with_mqa.input_quantizer TensorQuantizer(disabled) +model.layers.4.self_attn.kv_a_proj_with_mqa.output_quantizer TensorQuantizer(disabled) +model.layers.4.self_attn.kv_a_proj_with_mqa.weight_quantizer TensorQuantizer(disabled) +model.layers.4.self_attn.kv_b_proj.input_quantizer TensorQuantizer(disabled) +model.layers.4.self_attn.kv_b_proj.output_quantizer TensorQuantizer(disabled) +model.layers.4.self_attn.kv_b_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.4.self_attn.o_proj.input_quantizer TensorQuantizer(disabled) +model.layers.4.self_attn.o_proj.output_quantizer TensorQuantizer(disabled) +model.layers.4.self_attn.o_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.4.self_attn.q_bmm_quantizer TensorQuantizer(disabled) +model.layers.4.self_attn.k_bmm_quantizer TensorQuantizer((4, 3) bit fake per-tensor amax=4.48e+02(const) calibrator=MaxCalibrator quant) +model.layers.4.self_attn.v_bmm_quantizer TensorQuantizer((4, 3) bit fake per-tensor amax=4.48e+02(const) calibrator=MaxCalibrator quant) +model.layers.4.self_attn.p_bmm_quantizer TensorQuantizer(disabled) +model.layers.4.mlp.experts.gate_up_proj_input_quantizer TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.50e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.gate_up_proj_weight_quantizers.0 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.41e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.gate_up_proj_weight_quantizers.1 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.44e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.gate_up_proj_weight_quantizers.2 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.44e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.gate_up_proj_weight_quantizers.3 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.41e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.gate_up_proj_weight_quantizers.4 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.46e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.gate_up_proj_weight_quantizers.5 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.48e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.gate_up_proj_weight_quantizers.6 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.67e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.gate_up_proj_weight_quantizers.7 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.54e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.gate_up_proj_weight_quantizers.8 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.02e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.gate_up_proj_weight_quantizers.9 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.27e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.gate_up_proj_weight_quantizers.10 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.19e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.gate_up_proj_weight_quantizers.11 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.75e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.gate_up_proj_weight_quantizers.12 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.65e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.gate_up_proj_weight_quantizers.13 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.53e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.gate_up_proj_weight_quantizers.14 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.51e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.gate_up_proj_weight_quantizers.15 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.38e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.gate_up_proj_weight_quantizers.16 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.61e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.gate_up_proj_weight_quantizers.17 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.97e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.gate_up_proj_weight_quantizers.18 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.41e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.gate_up_proj_weight_quantizers.19 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.50e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.gate_up_proj_weight_quantizers.20 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.38e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.gate_up_proj_weight_quantizers.21 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.36e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.gate_up_proj_weight_quantizers.22 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.62e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.gate_up_proj_weight_quantizers.23 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.12e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.gate_up_proj_weight_quantizers.24 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.55e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.gate_up_proj_weight_quantizers.25 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.45e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.gate_up_proj_weight_quantizers.26 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.06e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.gate_up_proj_weight_quantizers.27 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.49e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.gate_up_proj_weight_quantizers.28 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.96e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.gate_up_proj_weight_quantizers.29 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.84e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.gate_up_proj_weight_quantizers.30 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.33e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.gate_up_proj_weight_quantizers.31 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.63e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.gate_up_proj_weight_quantizers.32 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.73e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.gate_up_proj_weight_quantizers.33 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.gate_up_proj_weight_quantizers.34 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.72e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.gate_up_proj_weight_quantizers.35 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.57e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.gate_up_proj_weight_quantizers.36 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.76e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.gate_up_proj_weight_quantizers.37 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.32e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.gate_up_proj_weight_quantizers.38 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.73e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.gate_up_proj_weight_quantizers.39 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.02e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.gate_up_proj_weight_quantizers.40 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.30e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.gate_up_proj_weight_quantizers.41 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.51e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.gate_up_proj_weight_quantizers.42 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.92e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.gate_up_proj_weight_quantizers.43 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.00e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.gate_up_proj_weight_quantizers.44 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.72e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.gate_up_proj_weight_quantizers.45 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.35e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.gate_up_proj_weight_quantizers.46 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.93e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.gate_up_proj_weight_quantizers.47 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.80e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.gate_up_proj_weight_quantizers.48 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.52e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.gate_up_proj_weight_quantizers.49 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.15e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.gate_up_proj_weight_quantizers.50 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.42e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.gate_up_proj_weight_quantizers.51 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.45e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.gate_up_proj_weight_quantizers.52 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.gate_up_proj_weight_quantizers.53 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.46e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.gate_up_proj_weight_quantizers.54 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.91e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.gate_up_proj_weight_quantizers.55 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.56e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.gate_up_proj_weight_quantizers.56 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.39e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.gate_up_proj_weight_quantizers.57 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.68e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.gate_up_proj_weight_quantizers.58 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.32e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.gate_up_proj_weight_quantizers.59 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.49e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.gate_up_proj_weight_quantizers.60 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.23e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.gate_up_proj_weight_quantizers.61 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.35e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.gate_up_proj_weight_quantizers.62 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.32e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.gate_up_proj_weight_quantizers.63 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.55e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.gate_up_proj_weight_quantizers.64 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.63e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.gate_up_proj_weight_quantizers.65 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.30e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.gate_up_proj_weight_quantizers.66 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.50e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.gate_up_proj_weight_quantizers.67 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.63e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.gate_up_proj_weight_quantizers.68 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.39e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.gate_up_proj_weight_quantizers.69 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.44e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.gate_up_proj_weight_quantizers.70 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.97e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.gate_up_proj_weight_quantizers.71 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.21e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.gate_up_proj_weight_quantizers.72 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.57e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.gate_up_proj_weight_quantizers.73 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.25e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.gate_up_proj_weight_quantizers.74 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.95e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.gate_up_proj_weight_quantizers.75 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.35e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.gate_up_proj_weight_quantizers.76 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.44e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.gate_up_proj_weight_quantizers.77 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.56e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.gate_up_proj_weight_quantizers.78 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.61e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.gate_up_proj_weight_quantizers.79 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.89e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.gate_up_proj_weight_quantizers.80 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.62e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.gate_up_proj_weight_quantizers.81 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.45e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.gate_up_proj_weight_quantizers.82 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.29e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.gate_up_proj_weight_quantizers.83 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.57e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.gate_up_proj_weight_quantizers.84 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.57e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.gate_up_proj_weight_quantizers.85 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.38e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.gate_up_proj_weight_quantizers.86 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.61e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.gate_up_proj_weight_quantizers.87 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.32e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.gate_up_proj_weight_quantizers.88 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.94e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.gate_up_proj_weight_quantizers.89 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.93e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.gate_up_proj_weight_quantizers.90 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.55e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.gate_up_proj_weight_quantizers.91 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.54e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.gate_up_proj_weight_quantizers.92 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.23e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.gate_up_proj_weight_quantizers.93 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.38e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.gate_up_proj_weight_quantizers.94 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.30e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.gate_up_proj_weight_quantizers.95 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.42e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.gate_up_proj_weight_quantizers.96 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.60e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.gate_up_proj_weight_quantizers.97 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.30e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.gate_up_proj_weight_quantizers.98 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.71e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.gate_up_proj_weight_quantizers.99 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.53e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.gate_up_proj_weight_quantizers.100 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.50e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.gate_up_proj_weight_quantizers.101 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.41e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.gate_up_proj_weight_quantizers.102 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.63e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.gate_up_proj_weight_quantizers.103 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.53e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.gate_up_proj_weight_quantizers.104 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.53e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.gate_up_proj_weight_quantizers.105 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.50e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.gate_up_proj_weight_quantizers.106 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.gate_up_proj_weight_quantizers.107 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.84e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.gate_up_proj_weight_quantizers.108 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.26e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.gate_up_proj_weight_quantizers.109 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.29e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.gate_up_proj_weight_quantizers.110 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.51e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.gate_up_proj_weight_quantizers.111 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.56e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.gate_up_proj_weight_quantizers.112 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.03e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.gate_up_proj_weight_quantizers.113 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.69e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.gate_up_proj_weight_quantizers.114 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.46e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.gate_up_proj_weight_quantizers.115 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.68e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.gate_up_proj_weight_quantizers.116 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.61e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.gate_up_proj_weight_quantizers.117 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.21e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.gate_up_proj_weight_quantizers.118 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.45e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.gate_up_proj_weight_quantizers.119 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.68e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.gate_up_proj_weight_quantizers.120 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.45e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.gate_up_proj_weight_quantizers.121 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.49e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.gate_up_proj_weight_quantizers.122 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.46e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.gate_up_proj_weight_quantizers.123 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.78e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.gate_up_proj_weight_quantizers.124 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.36e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.gate_up_proj_weight_quantizers.125 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.76e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.gate_up_proj_weight_quantizers.126 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.07e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.gate_up_proj_weight_quantizers.127 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.47e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.gate_up_proj_weight_quantizers.128 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.65e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.gate_up_proj_weight_quantizers.129 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.39e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.gate_up_proj_weight_quantizers.130 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.83e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.gate_up_proj_weight_quantizers.131 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.22e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.gate_up_proj_weight_quantizers.132 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.29e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.gate_up_proj_weight_quantizers.133 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.24e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.gate_up_proj_weight_quantizers.134 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.40e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.gate_up_proj_weight_quantizers.135 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.98e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.gate_up_proj_weight_quantizers.136 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.93e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.gate_up_proj_weight_quantizers.137 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.50e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.gate_up_proj_weight_quantizers.138 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.38e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.gate_up_proj_weight_quantizers.139 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.90e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.gate_up_proj_weight_quantizers.140 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.78e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.gate_up_proj_weight_quantizers.141 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.70e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.gate_up_proj_weight_quantizers.142 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.57e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.gate_up_proj_weight_quantizers.143 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.51e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.gate_up_proj_weight_quantizers.144 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.19e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.gate_up_proj_weight_quantizers.145 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.36e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.gate_up_proj_weight_quantizers.146 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.60e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.gate_up_proj_weight_quantizers.147 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.73e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.gate_up_proj_weight_quantizers.148 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.52e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.gate_up_proj_weight_quantizers.149 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.gate_up_proj_weight_quantizers.150 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.53e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.gate_up_proj_weight_quantizers.151 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.63e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.gate_up_proj_weight_quantizers.152 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.74e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.gate_up_proj_weight_quantizers.153 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.gate_up_proj_weight_quantizers.154 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.53e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.gate_up_proj_weight_quantizers.155 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.68e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.gate_up_proj_weight_quantizers.156 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.30e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.gate_up_proj_weight_quantizers.157 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.41e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.gate_up_proj_weight_quantizers.158 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.50e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.gate_up_proj_weight_quantizers.159 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.49e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.gate_up_proj_weight_quantizers.160 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.13e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.gate_up_proj_weight_quantizers.161 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.52e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.gate_up_proj_weight_quantizers.162 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.46e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.gate_up_proj_weight_quantizers.163 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.89e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.gate_up_proj_weight_quantizers.164 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.76e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.gate_up_proj_weight_quantizers.165 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.57e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.gate_up_proj_weight_quantizers.166 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.45e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.gate_up_proj_weight_quantizers.167 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.45e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.gate_up_proj_weight_quantizers.168 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.48e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.gate_up_proj_weight_quantizers.169 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.58e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.gate_up_proj_weight_quantizers.170 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.32e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.gate_up_proj_weight_quantizers.171 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.88e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.gate_up_proj_weight_quantizers.172 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.43e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.gate_up_proj_weight_quantizers.173 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.38e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.gate_up_proj_weight_quantizers.174 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.21e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.gate_up_proj_weight_quantizers.175 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.64e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.gate_up_proj_weight_quantizers.176 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.37e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.gate_up_proj_weight_quantizers.177 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.76e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.gate_up_proj_weight_quantizers.178 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.89e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.gate_up_proj_weight_quantizers.179 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.24e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.gate_up_proj_weight_quantizers.180 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.50e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.gate_up_proj_weight_quantizers.181 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.69e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.gate_up_proj_weight_quantizers.182 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.53e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.gate_up_proj_weight_quantizers.183 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.52e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.gate_up_proj_weight_quantizers.184 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.59e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.gate_up_proj_weight_quantizers.185 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.86e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.gate_up_proj_weight_quantizers.186 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.77e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.gate_up_proj_weight_quantizers.187 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.71e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.gate_up_proj_weight_quantizers.188 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.46e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.gate_up_proj_weight_quantizers.189 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.28e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.gate_up_proj_weight_quantizers.190 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.46e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.gate_up_proj_weight_quantizers.191 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.44e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.gate_up_proj_weight_quantizers.192 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.46e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.gate_up_proj_weight_quantizers.193 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.97e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.gate_up_proj_weight_quantizers.194 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.03e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.gate_up_proj_weight_quantizers.195 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.50e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.gate_up_proj_weight_quantizers.196 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.16e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.gate_up_proj_weight_quantizers.197 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.25e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.gate_up_proj_weight_quantizers.198 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.33e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.gate_up_proj_weight_quantizers.199 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.52e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.gate_up_proj_weight_quantizers.200 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.69e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.gate_up_proj_weight_quantizers.201 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.40e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.gate_up_proj_weight_quantizers.202 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.23e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.gate_up_proj_weight_quantizers.203 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.79e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.gate_up_proj_weight_quantizers.204 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.30e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.gate_up_proj_weight_quantizers.205 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.40e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.gate_up_proj_weight_quantizers.206 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.29e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.gate_up_proj_weight_quantizers.207 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.52e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.gate_up_proj_weight_quantizers.208 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.59e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.gate_up_proj_weight_quantizers.209 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.41e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.gate_up_proj_weight_quantizers.210 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.46e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.gate_up_proj_weight_quantizers.211 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.56e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.gate_up_proj_weight_quantizers.212 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.32e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.gate_up_proj_weight_quantizers.213 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.31e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.gate_up_proj_weight_quantizers.214 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.44e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.gate_up_proj_weight_quantizers.215 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.74e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.gate_up_proj_weight_quantizers.216 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.83e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.gate_up_proj_weight_quantizers.217 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.35e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.gate_up_proj_weight_quantizers.218 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.27e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.gate_up_proj_weight_quantizers.219 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.09e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.gate_up_proj_weight_quantizers.220 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.41e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.gate_up_proj_weight_quantizers.221 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.gate_up_proj_weight_quantizers.222 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.93e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.gate_up_proj_weight_quantizers.223 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.73e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.gate_up_proj_weight_quantizers.224 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.66e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.gate_up_proj_weight_quantizers.225 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.42e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.gate_up_proj_weight_quantizers.226 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.87e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.gate_up_proj_weight_quantizers.227 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.62e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.gate_up_proj_weight_quantizers.228 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.23e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.gate_up_proj_weight_quantizers.229 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.98e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.gate_up_proj_weight_quantizers.230 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.56e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.gate_up_proj_weight_quantizers.231 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.47e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.gate_up_proj_weight_quantizers.232 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.46e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.gate_up_proj_weight_quantizers.233 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.30e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.gate_up_proj_weight_quantizers.234 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.34e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.gate_up_proj_weight_quantizers.235 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.90e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.gate_up_proj_weight_quantizers.236 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.45e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.gate_up_proj_weight_quantizers.237 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.93e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.gate_up_proj_weight_quantizers.238 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.85e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.gate_up_proj_weight_quantizers.239 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.31e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.gate_up_proj_weight_quantizers.240 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.46e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.gate_up_proj_weight_quantizers.241 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.84e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.gate_up_proj_weight_quantizers.242 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.64e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.gate_up_proj_weight_quantizers.243 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.44e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.gate_up_proj_weight_quantizers.244 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.47e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.gate_up_proj_weight_quantizers.245 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.88e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.gate_up_proj_weight_quantizers.246 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.48e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.gate_up_proj_weight_quantizers.247 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.71e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.gate_up_proj_weight_quantizers.248 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.82e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.gate_up_proj_weight_quantizers.249 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.40e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.gate_up_proj_weight_quantizers.250 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.95e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.gate_up_proj_weight_quantizers.251 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.56e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.gate_up_proj_weight_quantizers.252 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.73e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.gate_up_proj_weight_quantizers.253 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.46e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.gate_up_proj_weight_quantizers.254 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.gate_up_proj_weight_quantizers.255 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.29e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.down_proj_input_quantizer TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.67e+00 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.down_proj_weight_quantizers.0 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.down_proj_weight_quantizers.1 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.down_proj_weight_quantizers.2 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.down_proj_weight_quantizers.3 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.42e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.down_proj_weight_quantizers.4 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.14e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.down_proj_weight_quantizers.5 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.down_proj_weight_quantizers.6 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.25e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.down_proj_weight_quantizers.7 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.down_proj_weight_quantizers.8 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.20e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.down_proj_weight_quantizers.9 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.37e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.down_proj_weight_quantizers.10 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.down_proj_weight_quantizers.11 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.32e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.down_proj_weight_quantizers.12 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.22e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.down_proj_weight_quantizers.13 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.41e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.down_proj_weight_quantizers.14 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.down_proj_weight_quantizers.15 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.62e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.down_proj_weight_quantizers.16 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.down_proj_weight_quantizers.17 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.65e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.down_proj_weight_quantizers.18 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.77e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.down_proj_weight_quantizers.19 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.down_proj_weight_quantizers.20 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.20e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.down_proj_weight_quantizers.21 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.down_proj_weight_quantizers.22 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.19e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.down_proj_weight_quantizers.23 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.73e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.down_proj_weight_quantizers.24 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.down_proj_weight_quantizers.25 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.23e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.down_proj_weight_quantizers.26 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.66e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.down_proj_weight_quantizers.27 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.down_proj_weight_quantizers.28 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.58e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.down_proj_weight_quantizers.29 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.15e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.down_proj_weight_quantizers.30 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.04e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.down_proj_weight_quantizers.31 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.44e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.down_proj_weight_quantizers.32 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.down_proj_weight_quantizers.33 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.41e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.down_proj_weight_quantizers.34 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.down_proj_weight_quantizers.35 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.18e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.down_proj_weight_quantizers.36 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.22e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.down_proj_weight_quantizers.37 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.down_proj_weight_quantizers.38 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.81e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.down_proj_weight_quantizers.39 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.31e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.down_proj_weight_quantizers.40 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.24e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.down_proj_weight_quantizers.41 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.30e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.down_proj_weight_quantizers.42 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.82e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.down_proj_weight_quantizers.43 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.down_proj_weight_quantizers.44 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.36e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.down_proj_weight_quantizers.45 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.down_proj_weight_quantizers.46 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.33e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.down_proj_weight_quantizers.47 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.75e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.down_proj_weight_quantizers.48 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.down_proj_weight_quantizers.49 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.22e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.down_proj_weight_quantizers.50 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.34e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.down_proj_weight_quantizers.51 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.down_proj_weight_quantizers.52 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.down_proj_weight_quantizers.53 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.down_proj_weight_quantizers.54 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.30e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.down_proj_weight_quantizers.55 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.down_proj_weight_quantizers.56 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.down_proj_weight_quantizers.57 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.50e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.down_proj_weight_quantizers.58 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.17e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.down_proj_weight_quantizers.59 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.down_proj_weight_quantizers.60 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.down_proj_weight_quantizers.61 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.29e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.down_proj_weight_quantizers.62 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.down_proj_weight_quantizers.63 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.35e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.down_proj_weight_quantizers.64 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.21e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.down_proj_weight_quantizers.65 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.down_proj_weight_quantizers.66 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.down_proj_weight_quantizers.67 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.14e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.down_proj_weight_quantizers.68 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.30e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.down_proj_weight_quantizers.69 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.87e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.down_proj_weight_quantizers.70 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.15e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.down_proj_weight_quantizers.71 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.20e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.down_proj_weight_quantizers.72 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.down_proj_weight_quantizers.73 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.down_proj_weight_quantizers.74 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.69e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.down_proj_weight_quantizers.75 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.23e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.down_proj_weight_quantizers.76 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.36e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.down_proj_weight_quantizers.77 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.79e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.down_proj_weight_quantizers.78 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.34e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.down_proj_weight_quantizers.79 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.29e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.down_proj_weight_quantizers.80 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.down_proj_weight_quantizers.81 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.down_proj_weight_quantizers.82 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.down_proj_weight_quantizers.83 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.23e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.down_proj_weight_quantizers.84 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.07e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.down_proj_weight_quantizers.85 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.15e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.down_proj_weight_quantizers.86 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.down_proj_weight_quantizers.87 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.34e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.down_proj_weight_quantizers.88 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.down_proj_weight_quantizers.89 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.39e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.down_proj_weight_quantizers.90 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.30e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.down_proj_weight_quantizers.91 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.down_proj_weight_quantizers.92 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.down_proj_weight_quantizers.93 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.35e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.down_proj_weight_quantizers.94 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.down_proj_weight_quantizers.95 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.down_proj_weight_quantizers.96 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.48e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.down_proj_weight_quantizers.97 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.down_proj_weight_quantizers.98 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.39e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.down_proj_weight_quantizers.99 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.47e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.down_proj_weight_quantizers.100 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.down_proj_weight_quantizers.101 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.down_proj_weight_quantizers.102 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.20e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.down_proj_weight_quantizers.103 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.down_proj_weight_quantizers.104 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.15e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.down_proj_weight_quantizers.105 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.33e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.down_proj_weight_quantizers.106 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.66e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.down_proj_weight_quantizers.107 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.down_proj_weight_quantizers.108 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.down_proj_weight_quantizers.109 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.22e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.down_proj_weight_quantizers.110 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.down_proj_weight_quantizers.111 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.down_proj_weight_quantizers.112 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.35e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.down_proj_weight_quantizers.113 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.18e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.down_proj_weight_quantizers.114 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.down_proj_weight_quantizers.115 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.41e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.down_proj_weight_quantizers.116 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.51e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.down_proj_weight_quantizers.117 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.down_proj_weight_quantizers.118 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.94e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.down_proj_weight_quantizers.119 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.down_proj_weight_quantizers.120 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.19e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.down_proj_weight_quantizers.121 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.59e-02 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.down_proj_weight_quantizers.122 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.down_proj_weight_quantizers.123 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.down_proj_weight_quantizers.124 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.72e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.down_proj_weight_quantizers.125 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.23e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.down_proj_weight_quantizers.126 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.85e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.down_proj_weight_quantizers.127 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.07e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.down_proj_weight_quantizers.128 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.75e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.down_proj_weight_quantizers.129 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.down_proj_weight_quantizers.130 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.down_proj_weight_quantizers.131 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.down_proj_weight_quantizers.132 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.down_proj_weight_quantizers.133 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.60e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.down_proj_weight_quantizers.134 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.55e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.down_proj_weight_quantizers.135 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.44e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.down_proj_weight_quantizers.136 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.62e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.down_proj_weight_quantizers.137 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.53e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.down_proj_weight_quantizers.138 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.14e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.down_proj_weight_quantizers.139 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.down_proj_weight_quantizers.140 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.50e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.down_proj_weight_quantizers.141 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.53e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.down_proj_weight_quantizers.142 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.down_proj_weight_quantizers.143 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.29e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.down_proj_weight_quantizers.144 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.20e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.down_proj_weight_quantizers.145 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.down_proj_weight_quantizers.146 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.50e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.down_proj_weight_quantizers.147 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.36e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.down_proj_weight_quantizers.148 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.13e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.down_proj_weight_quantizers.149 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.down_proj_weight_quantizers.150 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.05e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.down_proj_weight_quantizers.151 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.down_proj_weight_quantizers.152 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.92e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.down_proj_weight_quantizers.153 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.50e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.down_proj_weight_quantizers.154 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.84e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.down_proj_weight_quantizers.155 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.down_proj_weight_quantizers.156 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.down_proj_weight_quantizers.157 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.54e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.down_proj_weight_quantizers.158 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.31e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.down_proj_weight_quantizers.159 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.50e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.down_proj_weight_quantizers.160 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.25e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.down_proj_weight_quantizers.161 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.down_proj_weight_quantizers.162 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.43e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.down_proj_weight_quantizers.163 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.72e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.down_proj_weight_quantizers.164 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.down_proj_weight_quantizers.165 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.19e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.down_proj_weight_quantizers.166 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.37e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.down_proj_weight_quantizers.167 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.down_proj_weight_quantizers.168 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.68e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.down_proj_weight_quantizers.169 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.down_proj_weight_quantizers.170 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.down_proj_weight_quantizers.171 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.49e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.down_proj_weight_quantizers.172 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.down_proj_weight_quantizers.173 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.down_proj_weight_quantizers.174 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.27e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.down_proj_weight_quantizers.175 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.19e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.down_proj_weight_quantizers.176 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.down_proj_weight_quantizers.177 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.down_proj_weight_quantizers.178 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.48e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.down_proj_weight_quantizers.179 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.down_proj_weight_quantizers.180 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.19e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.down_proj_weight_quantizers.181 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.down_proj_weight_quantizers.182 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.down_proj_weight_quantizers.183 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.down_proj_weight_quantizers.184 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.89e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.down_proj_weight_quantizers.185 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.46e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.down_proj_weight_quantizers.186 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.26e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.down_proj_weight_quantizers.187 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.down_proj_weight_quantizers.188 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.down_proj_weight_quantizers.189 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.58e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.down_proj_weight_quantizers.190 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.down_proj_weight_quantizers.191 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.25e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.down_proj_weight_quantizers.192 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.down_proj_weight_quantizers.193 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.16e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.down_proj_weight_quantizers.194 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.23e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.down_proj_weight_quantizers.195 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.down_proj_weight_quantizers.196 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.25e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.down_proj_weight_quantizers.197 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.down_proj_weight_quantizers.198 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.down_proj_weight_quantizers.199 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.22e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.down_proj_weight_quantizers.200 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.19e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.down_proj_weight_quantizers.201 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.down_proj_weight_quantizers.202 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.down_proj_weight_quantizers.203 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.46e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.down_proj_weight_quantizers.204 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.down_proj_weight_quantizers.205 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.33e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.down_proj_weight_quantizers.206 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.03e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.down_proj_weight_quantizers.207 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.76e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.down_proj_weight_quantizers.208 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.down_proj_weight_quantizers.209 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.87e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.down_proj_weight_quantizers.210 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.55e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.down_proj_weight_quantizers.211 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.down_proj_weight_quantizers.212 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.13e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.down_proj_weight_quantizers.213 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.61e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.down_proj_weight_quantizers.214 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.62e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.down_proj_weight_quantizers.215 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.61e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.down_proj_weight_quantizers.216 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.43e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.down_proj_weight_quantizers.217 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.18e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.down_proj_weight_quantizers.218 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.down_proj_weight_quantizers.219 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.62e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.down_proj_weight_quantizers.220 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.down_proj_weight_quantizers.221 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.down_proj_weight_quantizers.222 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.44e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.down_proj_weight_quantizers.223 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.down_proj_weight_quantizers.224 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.38e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.down_proj_weight_quantizers.225 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.33e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.down_proj_weight_quantizers.226 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.68e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.down_proj_weight_quantizers.227 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.down_proj_weight_quantizers.228 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.down_proj_weight_quantizers.229 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.86e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.down_proj_weight_quantizers.230 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.39e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.down_proj_weight_quantizers.231 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.down_proj_weight_quantizers.232 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.down_proj_weight_quantizers.233 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.46e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.down_proj_weight_quantizers.234 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.43e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.down_proj_weight_quantizers.235 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.46e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.down_proj_weight_quantizers.236 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.down_proj_weight_quantizers.237 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.18e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.down_proj_weight_quantizers.238 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.down_proj_weight_quantizers.239 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.down_proj_weight_quantizers.240 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.down_proj_weight_quantizers.241 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.83e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.down_proj_weight_quantizers.242 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.33e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.down_proj_weight_quantizers.243 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.down_proj_weight_quantizers.244 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.32e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.down_proj_weight_quantizers.245 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.down_proj_weight_quantizers.246 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.29e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.down_proj_weight_quantizers.247 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.down_proj_weight_quantizers.248 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.33e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.down_proj_weight_quantizers.249 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.down_proj_weight_quantizers.250 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.47e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.down_proj_weight_quantizers.251 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.33e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.down_proj_weight_quantizers.252 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.down_proj_weight_quantizers.253 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.down_proj_weight_quantizers.254 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.down_proj_weight_quantizers.255 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.28e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.shared_experts.gate_proj.input_quantizer TensorQuantizer(disabled) +model.layers.4.mlp.shared_experts.gate_proj.output_quantizer TensorQuantizer(disabled) +model.layers.4.mlp.shared_experts.gate_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.4.mlp.shared_experts.up_proj.input_quantizer TensorQuantizer(disabled) +model.layers.4.mlp.shared_experts.up_proj.output_quantizer TensorQuantizer(disabled) +model.layers.4.mlp.shared_experts.up_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.4.mlp.shared_experts.down_proj.input_quantizer TensorQuantizer(disabled) +model.layers.4.mlp.shared_experts.down_proj.output_quantizer TensorQuantizer(disabled) +model.layers.4.mlp.shared_experts.down_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.5.self_attn.q_a_proj.input_quantizer TensorQuantizer(disabled) +model.layers.5.self_attn.q_a_proj.output_quantizer TensorQuantizer(disabled) +model.layers.5.self_attn.q_a_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.5.self_attn.q_b_proj.input_quantizer TensorQuantizer(disabled) +model.layers.5.self_attn.q_b_proj.output_quantizer TensorQuantizer(disabled) +model.layers.5.self_attn.q_b_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.5.self_attn.kv_a_proj_with_mqa.input_quantizer TensorQuantizer(disabled) +model.layers.5.self_attn.kv_a_proj_with_mqa.output_quantizer TensorQuantizer(disabled) +model.layers.5.self_attn.kv_a_proj_with_mqa.weight_quantizer TensorQuantizer(disabled) +model.layers.5.self_attn.kv_b_proj.input_quantizer TensorQuantizer(disabled) +model.layers.5.self_attn.kv_b_proj.output_quantizer TensorQuantizer(disabled) +model.layers.5.self_attn.kv_b_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.5.self_attn.o_proj.input_quantizer TensorQuantizer(disabled) +model.layers.5.self_attn.o_proj.output_quantizer TensorQuantizer(disabled) +model.layers.5.self_attn.o_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.5.self_attn.q_bmm_quantizer TensorQuantizer(disabled) +model.layers.5.self_attn.k_bmm_quantizer TensorQuantizer((4, 3) bit fake per-tensor amax=4.48e+02(const) calibrator=MaxCalibrator quant) +model.layers.5.self_attn.v_bmm_quantizer TensorQuantizer((4, 3) bit fake per-tensor amax=4.48e+02(const) calibrator=MaxCalibrator quant) +model.layers.5.self_attn.p_bmm_quantizer TensorQuantizer(disabled) +model.layers.5.mlp.experts.gate_up_proj_input_quantizer TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.73e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.gate_up_proj_weight_quantizers.0 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.38e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.gate_up_proj_weight_quantizers.1 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.22e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.gate_up_proj_weight_quantizers.2 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.44e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.gate_up_proj_weight_quantizers.3 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.29e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.gate_up_proj_weight_quantizers.4 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.81e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.gate_up_proj_weight_quantizers.5 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.33e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.gate_up_proj_weight_quantizers.6 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.78e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.gate_up_proj_weight_quantizers.7 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.60e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.gate_up_proj_weight_quantizers.8 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.19e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.gate_up_proj_weight_quantizers.9 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.gate_up_proj_weight_quantizers.10 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.14e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.gate_up_proj_weight_quantizers.11 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.59e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.gate_up_proj_weight_quantizers.12 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.06e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.gate_up_proj_weight_quantizers.13 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.32e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.gate_up_proj_weight_quantizers.14 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.38e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.gate_up_proj_weight_quantizers.15 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.59e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.gate_up_proj_weight_quantizers.16 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.97e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.gate_up_proj_weight_quantizers.17 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.21e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.gate_up_proj_weight_quantizers.18 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.70e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.gate_up_proj_weight_quantizers.19 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.67e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.gate_up_proj_weight_quantizers.20 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.13e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.gate_up_proj_weight_quantizers.21 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.50e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.gate_up_proj_weight_quantizers.22 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.87e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.gate_up_proj_weight_quantizers.23 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.89e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.gate_up_proj_weight_quantizers.24 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.82e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.gate_up_proj_weight_quantizers.25 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.46e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.gate_up_proj_weight_quantizers.26 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.37e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.gate_up_proj_weight_quantizers.27 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.57e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.gate_up_proj_weight_quantizers.28 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.67e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.gate_up_proj_weight_quantizers.29 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.81e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.gate_up_proj_weight_quantizers.30 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.43e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.gate_up_proj_weight_quantizers.31 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.14e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.gate_up_proj_weight_quantizers.32 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.70e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.gate_up_proj_weight_quantizers.33 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.73e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.gate_up_proj_weight_quantizers.34 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.56e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.gate_up_proj_weight_quantizers.35 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.28e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.gate_up_proj_weight_quantizers.36 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.62e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.gate_up_proj_weight_quantizers.37 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.33e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.gate_up_proj_weight_quantizers.38 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.56e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.gate_up_proj_weight_quantizers.39 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.87e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.gate_up_proj_weight_quantizers.40 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.70e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.gate_up_proj_weight_quantizers.41 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.63e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.gate_up_proj_weight_quantizers.42 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.37e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.gate_up_proj_weight_quantizers.43 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.50e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.gate_up_proj_weight_quantizers.44 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.82e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.gate_up_proj_weight_quantizers.45 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.28e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.gate_up_proj_weight_quantizers.46 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.75e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.gate_up_proj_weight_quantizers.47 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.46e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.gate_up_proj_weight_quantizers.48 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.16e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.gate_up_proj_weight_quantizers.49 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.56e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.gate_up_proj_weight_quantizers.50 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.54e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.gate_up_proj_weight_quantizers.51 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.34e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.gate_up_proj_weight_quantizers.52 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.63e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.gate_up_proj_weight_quantizers.53 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.82e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.gate_up_proj_weight_quantizers.54 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.62e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.gate_up_proj_weight_quantizers.55 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.23e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.gate_up_proj_weight_quantizers.56 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.58e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.gate_up_proj_weight_quantizers.57 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.47e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.gate_up_proj_weight_quantizers.58 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.70e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.gate_up_proj_weight_quantizers.59 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.48e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.gate_up_proj_weight_quantizers.60 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.53e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.gate_up_proj_weight_quantizers.61 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.32e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.gate_up_proj_weight_quantizers.62 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.77e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.gate_up_proj_weight_quantizers.63 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.80e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.gate_up_proj_weight_quantizers.64 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.34e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.gate_up_proj_weight_quantizers.65 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.57e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.gate_up_proj_weight_quantizers.66 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.33e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.gate_up_proj_weight_quantizers.67 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.75e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.gate_up_proj_weight_quantizers.68 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.52e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.gate_up_proj_weight_quantizers.69 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.76e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.gate_up_proj_weight_quantizers.70 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.58e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.gate_up_proj_weight_quantizers.71 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.96e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.gate_up_proj_weight_quantizers.72 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.60e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.gate_up_proj_weight_quantizers.73 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.60e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.gate_up_proj_weight_quantizers.74 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.87e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.gate_up_proj_weight_quantizers.75 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.77e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.gate_up_proj_weight_quantizers.76 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.47e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.gate_up_proj_weight_quantizers.77 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.67e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.gate_up_proj_weight_quantizers.78 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.50e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.gate_up_proj_weight_quantizers.79 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.30e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.gate_up_proj_weight_quantizers.80 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.56e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.gate_up_proj_weight_quantizers.81 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.05e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.gate_up_proj_weight_quantizers.82 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.48e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.gate_up_proj_weight_quantizers.83 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.65e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.gate_up_proj_weight_quantizers.84 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.02e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.gate_up_proj_weight_quantizers.85 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.47e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.gate_up_proj_weight_quantizers.86 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.31e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.gate_up_proj_weight_quantizers.87 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.49e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.gate_up_proj_weight_quantizers.88 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.30e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.gate_up_proj_weight_quantizers.89 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.68e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.gate_up_proj_weight_quantizers.90 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.31e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.gate_up_proj_weight_quantizers.91 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.89e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.gate_up_proj_weight_quantizers.92 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.58e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.gate_up_proj_weight_quantizers.93 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.58e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.gate_up_proj_weight_quantizers.94 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.61e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.gate_up_proj_weight_quantizers.95 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.55e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.gate_up_proj_weight_quantizers.96 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.57e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.gate_up_proj_weight_quantizers.97 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.70e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.gate_up_proj_weight_quantizers.98 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.83e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.gate_up_proj_weight_quantizers.99 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.33e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.gate_up_proj_weight_quantizers.100 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.88e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.gate_up_proj_weight_quantizers.101 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.52e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.gate_up_proj_weight_quantizers.102 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.67e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.gate_up_proj_weight_quantizers.103 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.22e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.gate_up_proj_weight_quantizers.104 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.74e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.gate_up_proj_weight_quantizers.105 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.34e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.gate_up_proj_weight_quantizers.106 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.42e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.gate_up_proj_weight_quantizers.107 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.48e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.gate_up_proj_weight_quantizers.108 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.33e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.gate_up_proj_weight_quantizers.109 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.49e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.gate_up_proj_weight_quantizers.110 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.35e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.gate_up_proj_weight_quantizers.111 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.32e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.gate_up_proj_weight_quantizers.112 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.83e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.gate_up_proj_weight_quantizers.113 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.85e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.gate_up_proj_weight_quantizers.114 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.92e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.gate_up_proj_weight_quantizers.115 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.94e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.gate_up_proj_weight_quantizers.116 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.67e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.gate_up_proj_weight_quantizers.117 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.52e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.gate_up_proj_weight_quantizers.118 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.77e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.gate_up_proj_weight_quantizers.119 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.27e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.gate_up_proj_weight_quantizers.120 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.48e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.gate_up_proj_weight_quantizers.121 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.86e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.gate_up_proj_weight_quantizers.122 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.37e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.gate_up_proj_weight_quantizers.123 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.67e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.gate_up_proj_weight_quantizers.124 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.76e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.gate_up_proj_weight_quantizers.125 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.27e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.gate_up_proj_weight_quantizers.126 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.28e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.gate_up_proj_weight_quantizers.127 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.15e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.gate_up_proj_weight_quantizers.128 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.54e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.gate_up_proj_weight_quantizers.129 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.60e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.gate_up_proj_weight_quantizers.130 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.74e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.gate_up_proj_weight_quantizers.131 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.49e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.gate_up_proj_weight_quantizers.132 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.46e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.gate_up_proj_weight_quantizers.133 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.74e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.gate_up_proj_weight_quantizers.134 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.51e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.gate_up_proj_weight_quantizers.135 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.77e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.gate_up_proj_weight_quantizers.136 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.74e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.gate_up_proj_weight_quantizers.137 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.92e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.gate_up_proj_weight_quantizers.138 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.46e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.gate_up_proj_weight_quantizers.139 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.51e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.gate_up_proj_weight_quantizers.140 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.03e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.gate_up_proj_weight_quantizers.141 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.23e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.gate_up_proj_weight_quantizers.142 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.23e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.gate_up_proj_weight_quantizers.143 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.56e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.gate_up_proj_weight_quantizers.144 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.65e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.gate_up_proj_weight_quantizers.145 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.47e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.gate_up_proj_weight_quantizers.146 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.37e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.gate_up_proj_weight_quantizers.147 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=4.67e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.gate_up_proj_weight_quantizers.148 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.39e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.gate_up_proj_weight_quantizers.149 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.31e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.gate_up_proj_weight_quantizers.150 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.65e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.gate_up_proj_weight_quantizers.151 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.64e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.gate_up_proj_weight_quantizers.152 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.47e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.gate_up_proj_weight_quantizers.153 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.81e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.gate_up_proj_weight_quantizers.154 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.41e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.gate_up_proj_weight_quantizers.155 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.28e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.gate_up_proj_weight_quantizers.156 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.76e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.gate_up_proj_weight_quantizers.157 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.57e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.gate_up_proj_weight_quantizers.158 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.46e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.gate_up_proj_weight_quantizers.159 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.64e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.gate_up_proj_weight_quantizers.160 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.68e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.gate_up_proj_weight_quantizers.161 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.99e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.gate_up_proj_weight_quantizers.162 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.34e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.gate_up_proj_weight_quantizers.163 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.79e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.gate_up_proj_weight_quantizers.164 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.58e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.gate_up_proj_weight_quantizers.165 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.43e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.gate_up_proj_weight_quantizers.166 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.58e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.gate_up_proj_weight_quantizers.167 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.88e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.gate_up_proj_weight_quantizers.168 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.53e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.gate_up_proj_weight_quantizers.169 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.82e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.gate_up_proj_weight_quantizers.170 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.67e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.gate_up_proj_weight_quantizers.171 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.91e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.gate_up_proj_weight_quantizers.172 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.23e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.gate_up_proj_weight_quantizers.173 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.53e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.gate_up_proj_weight_quantizers.174 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.37e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.gate_up_proj_weight_quantizers.175 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.04e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.gate_up_proj_weight_quantizers.176 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.39e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.gate_up_proj_weight_quantizers.177 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.06e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.gate_up_proj_weight_quantizers.178 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.00e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.gate_up_proj_weight_quantizers.179 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.82e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.gate_up_proj_weight_quantizers.180 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.74e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.gate_up_proj_weight_quantizers.181 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.34e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.gate_up_proj_weight_quantizers.182 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.56e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.gate_up_proj_weight_quantizers.183 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.06e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.gate_up_proj_weight_quantizers.184 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.97e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.gate_up_proj_weight_quantizers.185 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.47e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.gate_up_proj_weight_quantizers.186 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.50e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.gate_up_proj_weight_quantizers.187 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.32e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.gate_up_proj_weight_quantizers.188 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.77e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.gate_up_proj_weight_quantizers.189 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.53e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.gate_up_proj_weight_quantizers.190 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.74e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.gate_up_proj_weight_quantizers.191 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.38e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.gate_up_proj_weight_quantizers.192 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.55e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.gate_up_proj_weight_quantizers.193 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.39e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.gate_up_proj_weight_quantizers.194 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.17e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.gate_up_proj_weight_quantizers.195 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.84e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.gate_up_proj_weight_quantizers.196 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.70e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.gate_up_proj_weight_quantizers.197 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.38e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.gate_up_proj_weight_quantizers.198 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.50e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.gate_up_proj_weight_quantizers.199 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.46e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.gate_up_proj_weight_quantizers.200 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.93e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.gate_up_proj_weight_quantizers.201 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.40e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.gate_up_proj_weight_quantizers.202 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.75e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.gate_up_proj_weight_quantizers.203 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.46e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.gate_up_proj_weight_quantizers.204 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.57e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.gate_up_proj_weight_quantizers.205 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.56e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.gate_up_proj_weight_quantizers.206 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.59e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.gate_up_proj_weight_quantizers.207 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.90e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.gate_up_proj_weight_quantizers.208 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.94e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.gate_up_proj_weight_quantizers.209 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.58e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.gate_up_proj_weight_quantizers.210 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.48e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.gate_up_proj_weight_quantizers.211 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.31e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.gate_up_proj_weight_quantizers.212 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.66e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.gate_up_proj_weight_quantizers.213 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.30e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.gate_up_proj_weight_quantizers.214 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.28e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.gate_up_proj_weight_quantizers.215 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.37e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.gate_up_proj_weight_quantizers.216 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.46e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.gate_up_proj_weight_quantizers.217 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.60e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.gate_up_proj_weight_quantizers.218 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.71e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.gate_up_proj_weight_quantizers.219 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.56e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.gate_up_proj_weight_quantizers.220 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.45e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.gate_up_proj_weight_quantizers.221 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.77e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.gate_up_proj_weight_quantizers.222 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.54e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.gate_up_proj_weight_quantizers.223 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.61e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.gate_up_proj_weight_quantizers.224 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.59e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.gate_up_proj_weight_quantizers.225 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.90e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.gate_up_proj_weight_quantizers.226 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.47e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.gate_up_proj_weight_quantizers.227 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.54e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.gate_up_proj_weight_quantizers.228 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.50e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.gate_up_proj_weight_quantizers.229 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.08e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.gate_up_proj_weight_quantizers.230 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.42e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.gate_up_proj_weight_quantizers.231 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.32e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.gate_up_proj_weight_quantizers.232 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.73e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.gate_up_proj_weight_quantizers.233 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.gate_up_proj_weight_quantizers.234 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.56e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.gate_up_proj_weight_quantizers.235 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.71e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.gate_up_proj_weight_quantizers.236 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.77e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.gate_up_proj_weight_quantizers.237 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.66e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.gate_up_proj_weight_quantizers.238 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.56e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.gate_up_proj_weight_quantizers.239 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.40e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.gate_up_proj_weight_quantizers.240 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.46e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.gate_up_proj_weight_quantizers.241 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.93e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.gate_up_proj_weight_quantizers.242 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.99e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.gate_up_proj_weight_quantizers.243 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.26e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.gate_up_proj_weight_quantizers.244 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.30e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.gate_up_proj_weight_quantizers.245 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.29e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.gate_up_proj_weight_quantizers.246 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.46e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.gate_up_proj_weight_quantizers.247 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.14e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.gate_up_proj_weight_quantizers.248 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.52e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.gate_up_proj_weight_quantizers.249 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.88e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.gate_up_proj_weight_quantizers.250 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.46e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.gate_up_proj_weight_quantizers.251 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.57e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.gate_up_proj_weight_quantizers.252 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.52e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.gate_up_proj_weight_quantizers.253 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.72e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.gate_up_proj_weight_quantizers.254 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.59e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.gate_up_proj_weight_quantizers.255 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.40e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.down_proj_input_quantizer TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.48e+00 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.down_proj_weight_quantizers.0 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.down_proj_weight_quantizers.1 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.down_proj_weight_quantizers.2 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.down_proj_weight_quantizers.3 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.17e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.down_proj_weight_quantizers.4 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.20e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.down_proj_weight_quantizers.5 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.77e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.down_proj_weight_quantizers.6 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.24e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.down_proj_weight_quantizers.7 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.down_proj_weight_quantizers.8 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.20e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.down_proj_weight_quantizers.9 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.64e-02 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.down_proj_weight_quantizers.10 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.down_proj_weight_quantizers.11 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.53e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.down_proj_weight_quantizers.12 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.down_proj_weight_quantizers.13 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.24e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.down_proj_weight_quantizers.14 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.25e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.down_proj_weight_quantizers.15 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.down_proj_weight_quantizers.16 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.down_proj_weight_quantizers.17 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.down_proj_weight_quantizers.18 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.06e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.down_proj_weight_quantizers.19 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.89e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.down_proj_weight_quantizers.20 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.down_proj_weight_quantizers.21 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.44e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.down_proj_weight_quantizers.22 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.36e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.down_proj_weight_quantizers.23 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.66e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.down_proj_weight_quantizers.24 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.15e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.down_proj_weight_quantizers.25 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.down_proj_weight_quantizers.26 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.down_proj_weight_quantizers.27 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.down_proj_weight_quantizers.28 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.36e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.down_proj_weight_quantizers.29 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.48e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.down_proj_weight_quantizers.30 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.17e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.down_proj_weight_quantizers.31 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.down_proj_weight_quantizers.32 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.18e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.down_proj_weight_quantizers.33 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.21e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.down_proj_weight_quantizers.34 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.59e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.down_proj_weight_quantizers.35 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.14e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.down_proj_weight_quantizers.36 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.63e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.down_proj_weight_quantizers.37 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.down_proj_weight_quantizers.38 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.down_proj_weight_quantizers.39 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.30e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.down_proj_weight_quantizers.40 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.43e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.down_proj_weight_quantizers.41 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.58e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.down_proj_weight_quantizers.42 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.26e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.down_proj_weight_quantizers.43 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.down_proj_weight_quantizers.44 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.23e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.down_proj_weight_quantizers.45 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.down_proj_weight_quantizers.46 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.74e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.down_proj_weight_quantizers.47 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.11e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.down_proj_weight_quantizers.48 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.down_proj_weight_quantizers.49 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.20e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.down_proj_weight_quantizers.50 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.down_proj_weight_quantizers.51 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.12e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.down_proj_weight_quantizers.52 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.29e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.down_proj_weight_quantizers.53 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.down_proj_weight_quantizers.54 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.40e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.down_proj_weight_quantizers.55 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.down_proj_weight_quantizers.56 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.down_proj_weight_quantizers.57 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.25e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.down_proj_weight_quantizers.58 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.69e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.down_proj_weight_quantizers.59 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.down_proj_weight_quantizers.60 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.down_proj_weight_quantizers.61 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.36e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.down_proj_weight_quantizers.62 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.92e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.down_proj_weight_quantizers.63 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.18e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.down_proj_weight_quantizers.64 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.down_proj_weight_quantizers.65 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.27e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.down_proj_weight_quantizers.66 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.43e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.down_proj_weight_quantizers.67 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.17e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.down_proj_weight_quantizers.68 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.99e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.down_proj_weight_quantizers.69 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.down_proj_weight_quantizers.70 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.44e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.down_proj_weight_quantizers.71 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.34e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.down_proj_weight_quantizers.72 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.50e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.down_proj_weight_quantizers.73 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.down_proj_weight_quantizers.74 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.75e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.down_proj_weight_quantizers.75 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.down_proj_weight_quantizers.76 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.down_proj_weight_quantizers.77 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.42e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.down_proj_weight_quantizers.78 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.down_proj_weight_quantizers.79 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.down_proj_weight_quantizers.80 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.13e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.down_proj_weight_quantizers.81 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.61e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.down_proj_weight_quantizers.82 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.88e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.down_proj_weight_quantizers.83 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.48e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.down_proj_weight_quantizers.84 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.00e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.down_proj_weight_quantizers.85 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.34e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.down_proj_weight_quantizers.86 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.21e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.down_proj_weight_quantizers.87 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.down_proj_weight_quantizers.88 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.down_proj_weight_quantizers.89 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.24e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.down_proj_weight_quantizers.90 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.down_proj_weight_quantizers.91 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.38e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.down_proj_weight_quantizers.92 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.42e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.down_proj_weight_quantizers.93 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.40e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.down_proj_weight_quantizers.94 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.43e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.down_proj_weight_quantizers.95 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.83e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.down_proj_weight_quantizers.96 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.down_proj_weight_quantizers.97 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.58e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.down_proj_weight_quantizers.98 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.down_proj_weight_quantizers.99 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.19e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.down_proj_weight_quantizers.100 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.down_proj_weight_quantizers.101 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.down_proj_weight_quantizers.102 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.23e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.down_proj_weight_quantizers.103 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.14e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.down_proj_weight_quantizers.104 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.21e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.down_proj_weight_quantizers.105 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.23e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.down_proj_weight_quantizers.106 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.down_proj_weight_quantizers.107 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.43e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.down_proj_weight_quantizers.108 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.13e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.down_proj_weight_quantizers.109 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.down_proj_weight_quantizers.110 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.down_proj_weight_quantizers.111 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.down_proj_weight_quantizers.112 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.09e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.down_proj_weight_quantizers.113 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.61e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.down_proj_weight_quantizers.114 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.71e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.down_proj_weight_quantizers.115 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.38e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.down_proj_weight_quantizers.116 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.39e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.down_proj_weight_quantizers.117 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.down_proj_weight_quantizers.118 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.04e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.down_proj_weight_quantizers.119 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.down_proj_weight_quantizers.120 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.down_proj_weight_quantizers.121 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.down_proj_weight_quantizers.122 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.42e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.down_proj_weight_quantizers.123 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.36e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.down_proj_weight_quantizers.124 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.down_proj_weight_quantizers.125 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.20e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.down_proj_weight_quantizers.126 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.23e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.down_proj_weight_quantizers.127 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.38e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.down_proj_weight_quantizers.128 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.29e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.down_proj_weight_quantizers.129 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.down_proj_weight_quantizers.130 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.22e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.down_proj_weight_quantizers.131 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.down_proj_weight_quantizers.132 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.down_proj_weight_quantizers.133 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.down_proj_weight_quantizers.134 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.51e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.down_proj_weight_quantizers.135 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.77e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.down_proj_weight_quantizers.136 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.55e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.down_proj_weight_quantizers.137 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.73e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.down_proj_weight_quantizers.138 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.72e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.down_proj_weight_quantizers.139 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.42e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.down_proj_weight_quantizers.140 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.48e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.down_proj_weight_quantizers.141 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.12e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.down_proj_weight_quantizers.142 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.down_proj_weight_quantizers.143 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.21e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.down_proj_weight_quantizers.144 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.47e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.down_proj_weight_quantizers.145 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.down_proj_weight_quantizers.146 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.46e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.down_proj_weight_quantizers.147 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=6.60e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.down_proj_weight_quantizers.148 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.down_proj_weight_quantizers.149 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.down_proj_weight_quantizers.150 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.47e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.down_proj_weight_quantizers.151 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.72e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.down_proj_weight_quantizers.152 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.27e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.down_proj_weight_quantizers.153 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.90e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.down_proj_weight_quantizers.154 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.20e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.down_proj_weight_quantizers.155 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.down_proj_weight_quantizers.156 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.64e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.down_proj_weight_quantizers.157 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.down_proj_weight_quantizers.158 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.down_proj_weight_quantizers.159 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.down_proj_weight_quantizers.160 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=4.55e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.down_proj_weight_quantizers.161 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.down_proj_weight_quantizers.162 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.31e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.down_proj_weight_quantizers.163 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.23e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.down_proj_weight_quantizers.164 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.25e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.down_proj_weight_quantizers.165 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.down_proj_weight_quantizers.166 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.down_proj_weight_quantizers.167 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.54e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.down_proj_weight_quantizers.168 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.27e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.down_proj_weight_quantizers.169 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.down_proj_weight_quantizers.170 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.39e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.down_proj_weight_quantizers.171 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.down_proj_weight_quantizers.172 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.51e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.down_proj_weight_quantizers.173 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.14e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.down_proj_weight_quantizers.174 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.46e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.down_proj_weight_quantizers.175 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.66e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.down_proj_weight_quantizers.176 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.down_proj_weight_quantizers.177 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.52e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.down_proj_weight_quantizers.178 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.down_proj_weight_quantizers.179 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.down_proj_weight_quantizers.180 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.down_proj_weight_quantizers.181 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.down_proj_weight_quantizers.182 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.64e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.down_proj_weight_quantizers.183 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.43e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.down_proj_weight_quantizers.184 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.17e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.down_proj_weight_quantizers.185 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.42e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.down_proj_weight_quantizers.186 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.down_proj_weight_quantizers.187 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.12e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.down_proj_weight_quantizers.188 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.56e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.down_proj_weight_quantizers.189 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.74e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.down_proj_weight_quantizers.190 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.46e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.down_proj_weight_quantizers.191 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.down_proj_weight_quantizers.192 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.down_proj_weight_quantizers.193 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.26e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.down_proj_weight_quantizers.194 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.18e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.down_proj_weight_quantizers.195 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.22e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.down_proj_weight_quantizers.196 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.65e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.down_proj_weight_quantizers.197 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.down_proj_weight_quantizers.198 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.60e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.down_proj_weight_quantizers.199 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.77e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.down_proj_weight_quantizers.200 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.down_proj_weight_quantizers.201 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.down_proj_weight_quantizers.202 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.48e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.down_proj_weight_quantizers.203 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.down_proj_weight_quantizers.204 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.26e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.down_proj_weight_quantizers.205 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.down_proj_weight_quantizers.206 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.down_proj_weight_quantizers.207 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.39e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.down_proj_weight_quantizers.208 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.98e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.down_proj_weight_quantizers.209 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.down_proj_weight_quantizers.210 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.34e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.down_proj_weight_quantizers.211 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.down_proj_weight_quantizers.212 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.down_proj_weight_quantizers.213 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.down_proj_weight_quantizers.214 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.down_proj_weight_quantizers.215 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.down_proj_weight_quantizers.216 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.down_proj_weight_quantizers.217 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.down_proj_weight_quantizers.218 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.17e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.down_proj_weight_quantizers.219 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.13e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.down_proj_weight_quantizers.220 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.down_proj_weight_quantizers.221 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.21e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.down_proj_weight_quantizers.222 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.92e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.down_proj_weight_quantizers.223 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.53e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.down_proj_weight_quantizers.224 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.56e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.down_proj_weight_quantizers.225 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.14e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.down_proj_weight_quantizers.226 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.down_proj_weight_quantizers.227 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.10e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.down_proj_weight_quantizers.228 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.down_proj_weight_quantizers.229 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.58e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.down_proj_weight_quantizers.230 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.down_proj_weight_quantizers.231 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.down_proj_weight_quantizers.232 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.03e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.down_proj_weight_quantizers.233 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.64e-02 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.down_proj_weight_quantizers.234 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.83e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.down_proj_weight_quantizers.235 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.down_proj_weight_quantizers.236 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.57e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.down_proj_weight_quantizers.237 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.down_proj_weight_quantizers.238 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.02e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.down_proj_weight_quantizers.239 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.down_proj_weight_quantizers.240 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.down_proj_weight_quantizers.241 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.25e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.down_proj_weight_quantizers.242 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.88e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.down_proj_weight_quantizers.243 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.down_proj_weight_quantizers.244 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.down_proj_weight_quantizers.245 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.down_proj_weight_quantizers.246 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.24e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.down_proj_weight_quantizers.247 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.93e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.down_proj_weight_quantizers.248 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.down_proj_weight_quantizers.249 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.46e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.down_proj_weight_quantizers.250 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.down_proj_weight_quantizers.251 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.74e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.down_proj_weight_quantizers.252 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.39e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.down_proj_weight_quantizers.253 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.26e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.down_proj_weight_quantizers.254 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.down_proj_weight_quantizers.255 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=4.39e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.shared_experts.gate_proj.input_quantizer TensorQuantizer(disabled) +model.layers.5.mlp.shared_experts.gate_proj.output_quantizer TensorQuantizer(disabled) +model.layers.5.mlp.shared_experts.gate_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.5.mlp.shared_experts.up_proj.input_quantizer TensorQuantizer(disabled) +model.layers.5.mlp.shared_experts.up_proj.output_quantizer TensorQuantizer(disabled) +model.layers.5.mlp.shared_experts.up_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.5.mlp.shared_experts.down_proj.input_quantizer TensorQuantizer(disabled) +model.layers.5.mlp.shared_experts.down_proj.output_quantizer TensorQuantizer(disabled) +model.layers.5.mlp.shared_experts.down_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.6.self_attn.q_a_proj.input_quantizer TensorQuantizer(disabled) +model.layers.6.self_attn.q_a_proj.output_quantizer TensorQuantizer(disabled) +model.layers.6.self_attn.q_a_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.6.self_attn.q_b_proj.input_quantizer TensorQuantizer(disabled) +model.layers.6.self_attn.q_b_proj.output_quantizer TensorQuantizer(disabled) +model.layers.6.self_attn.q_b_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.6.self_attn.kv_a_proj_with_mqa.input_quantizer TensorQuantizer(disabled) +model.layers.6.self_attn.kv_a_proj_with_mqa.output_quantizer TensorQuantizer(disabled) +model.layers.6.self_attn.kv_a_proj_with_mqa.weight_quantizer TensorQuantizer(disabled) +model.layers.6.self_attn.kv_b_proj.input_quantizer TensorQuantizer(disabled) +model.layers.6.self_attn.kv_b_proj.output_quantizer TensorQuantizer(disabled) +model.layers.6.self_attn.kv_b_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.6.self_attn.o_proj.input_quantizer TensorQuantizer(disabled) +model.layers.6.self_attn.o_proj.output_quantizer TensorQuantizer(disabled) +model.layers.6.self_attn.o_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.6.self_attn.indexer.wq_b.input_quantizer TensorQuantizer(disabled) +model.layers.6.self_attn.indexer.wq_b.output_quantizer TensorQuantizer(disabled) +model.layers.6.self_attn.indexer.wq_b.weight_quantizer TensorQuantizer(disabled) +model.layers.6.self_attn.indexer.wk.input_quantizer TensorQuantizer(disabled) +model.layers.6.self_attn.indexer.wk.output_quantizer TensorQuantizer(disabled) +model.layers.6.self_attn.indexer.wk.weight_quantizer TensorQuantizer(disabled) +model.layers.6.self_attn.indexer.k_norm.input_quantizer TensorQuantizer(disabled) +model.layers.6.self_attn.indexer.k_norm.output_quantizer TensorQuantizer(disabled) +model.layers.6.self_attn.indexer.weights_proj.input_quantizer TensorQuantizer(disabled) +model.layers.6.self_attn.indexer.weights_proj.output_quantizer TensorQuantizer(disabled) +model.layers.6.self_attn.indexer.weights_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.6.self_attn.q_bmm_quantizer TensorQuantizer(disabled) +model.layers.6.self_attn.k_bmm_quantizer TensorQuantizer((4, 3) bit fake per-tensor amax=4.48e+02(const) calibrator=MaxCalibrator quant) +model.layers.6.self_attn.v_bmm_quantizer TensorQuantizer((4, 3) bit fake per-tensor amax=4.48e+02(const) calibrator=MaxCalibrator quant) +model.layers.6.self_attn.p_bmm_quantizer TensorQuantizer(disabled) +model.layers.6.mlp.experts.gate_up_proj_input_quantizer TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.81e+00 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.gate_up_proj_weight_quantizers.0 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.15e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.gate_up_proj_weight_quantizers.1 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.35e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.gate_up_proj_weight_quantizers.2 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.42e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.gate_up_proj_weight_quantizers.3 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.15e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.gate_up_proj_weight_quantizers.4 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.32e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.gate_up_proj_weight_quantizers.5 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.15e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.gate_up_proj_weight_quantizers.6 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.45e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.gate_up_proj_weight_quantizers.7 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.23e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.gate_up_proj_weight_quantizers.8 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.27e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.gate_up_proj_weight_quantizers.9 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.89e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.gate_up_proj_weight_quantizers.10 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.31e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.gate_up_proj_weight_quantizers.11 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.28e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.gate_up_proj_weight_quantizers.12 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.35e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.gate_up_proj_weight_quantizers.13 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.28e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.gate_up_proj_weight_quantizers.14 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.37e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.gate_up_proj_weight_quantizers.15 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.31e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.gate_up_proj_weight_quantizers.16 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.30e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.gate_up_proj_weight_quantizers.17 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.26e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.gate_up_proj_weight_quantizers.18 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.gate_up_proj_weight_quantizers.19 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.gate_up_proj_weight_quantizers.20 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.29e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.gate_up_proj_weight_quantizers.21 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.09e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.gate_up_proj_weight_quantizers.22 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.30e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.gate_up_proj_weight_quantizers.23 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.25e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.gate_up_proj_weight_quantizers.24 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.36e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.gate_up_proj_weight_quantizers.25 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.18e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.gate_up_proj_weight_quantizers.26 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.21e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.gate_up_proj_weight_quantizers.27 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.15e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.gate_up_proj_weight_quantizers.28 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.gate_up_proj_weight_quantizers.29 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.65e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.gate_up_proj_weight_quantizers.30 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.25e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.gate_up_proj_weight_quantizers.31 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.gate_up_proj_weight_quantizers.32 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.15e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.gate_up_proj_weight_quantizers.33 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.gate_up_proj_weight_quantizers.34 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.53e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.gate_up_proj_weight_quantizers.35 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.46e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.gate_up_proj_weight_quantizers.36 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.40e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.gate_up_proj_weight_quantizers.37 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.74e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.gate_up_proj_weight_quantizers.38 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.gate_up_proj_weight_quantizers.39 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.27e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.gate_up_proj_weight_quantizers.40 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.36e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.gate_up_proj_weight_quantizers.41 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.34e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.gate_up_proj_weight_quantizers.42 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.22e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.gate_up_proj_weight_quantizers.43 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.gate_up_proj_weight_quantizers.44 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.gate_up_proj_weight_quantizers.45 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.27e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.gate_up_proj_weight_quantizers.46 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.19e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.gate_up_proj_weight_quantizers.47 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.45e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.gate_up_proj_weight_quantizers.48 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.26e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.gate_up_proj_weight_quantizers.49 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.14e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.gate_up_proj_weight_quantizers.50 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.21e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.gate_up_proj_weight_quantizers.51 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.18e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.gate_up_proj_weight_quantizers.52 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.18e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.gate_up_proj_weight_quantizers.53 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.13e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.gate_up_proj_weight_quantizers.54 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.71e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.gate_up_proj_weight_quantizers.55 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.gate_up_proj_weight_quantizers.56 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.33e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.gate_up_proj_weight_quantizers.57 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.gate_up_proj_weight_quantizers.58 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.72e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.gate_up_proj_weight_quantizers.59 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.35e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.gate_up_proj_weight_quantizers.60 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.73e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.gate_up_proj_weight_quantizers.61 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.gate_up_proj_weight_quantizers.62 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.79e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.gate_up_proj_weight_quantizers.63 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.44e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.gate_up_proj_weight_quantizers.64 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=5.74e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.gate_up_proj_weight_quantizers.65 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.33e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.gate_up_proj_weight_quantizers.66 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.19e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.gate_up_proj_weight_quantizers.67 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.52e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.gate_up_proj_weight_quantizers.68 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.14e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.gate_up_proj_weight_quantizers.69 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.21e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.gate_up_proj_weight_quantizers.70 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.14e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.gate_up_proj_weight_quantizers.71 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.26e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.gate_up_proj_weight_quantizers.72 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.14e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.gate_up_proj_weight_quantizers.73 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.30e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.gate_up_proj_weight_quantizers.74 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.22e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.gate_up_proj_weight_quantizers.75 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.20e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.gate_up_proj_weight_quantizers.76 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.48e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.gate_up_proj_weight_quantizers.77 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.31e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.gate_up_proj_weight_quantizers.78 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.30e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.gate_up_proj_weight_quantizers.79 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.gate_up_proj_weight_quantizers.80 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.14e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.gate_up_proj_weight_quantizers.81 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.44e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.gate_up_proj_weight_quantizers.82 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.gate_up_proj_weight_quantizers.83 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.gate_up_proj_weight_quantizers.84 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.47e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.gate_up_proj_weight_quantizers.85 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.13e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.gate_up_proj_weight_quantizers.86 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.gate_up_proj_weight_quantizers.87 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.32e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.gate_up_proj_weight_quantizers.88 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.31e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.gate_up_proj_weight_quantizers.89 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.23e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.gate_up_proj_weight_quantizers.90 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.gate_up_proj_weight_quantizers.91 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.37e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.gate_up_proj_weight_quantizers.92 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.45e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.gate_up_proj_weight_quantizers.93 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.42e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.gate_up_proj_weight_quantizers.94 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.22e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.gate_up_proj_weight_quantizers.95 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.36e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.gate_up_proj_weight_quantizers.96 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.22e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.gate_up_proj_weight_quantizers.97 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.27e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.gate_up_proj_weight_quantizers.98 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.15e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.gate_up_proj_weight_quantizers.99 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.45e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.gate_up_proj_weight_quantizers.100 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.32e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.gate_up_proj_weight_quantizers.101 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=4.22e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.gate_up_proj_weight_quantizers.102 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.gate_up_proj_weight_quantizers.103 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.17e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.gate_up_proj_weight_quantizers.104 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.45e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.gate_up_proj_weight_quantizers.105 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.25e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.gate_up_proj_weight_quantizers.106 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.22e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.gate_up_proj_weight_quantizers.107 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.26e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.gate_up_proj_weight_quantizers.108 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.19e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.gate_up_proj_weight_quantizers.109 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.15e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.gate_up_proj_weight_quantizers.110 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.16e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.gate_up_proj_weight_quantizers.111 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.gate_up_proj_weight_quantizers.112 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.23e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.gate_up_proj_weight_quantizers.113 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.gate_up_proj_weight_quantizers.114 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.gate_up_proj_weight_quantizers.115 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.43e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.gate_up_proj_weight_quantizers.116 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.62e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.gate_up_proj_weight_quantizers.117 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.31e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.gate_up_proj_weight_quantizers.118 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.gate_up_proj_weight_quantizers.119 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.57e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.gate_up_proj_weight_quantizers.120 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.73e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.gate_up_proj_weight_quantizers.121 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.27e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.gate_up_proj_weight_quantizers.122 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.46e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.gate_up_proj_weight_quantizers.123 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.27e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.gate_up_proj_weight_quantizers.124 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.42e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.gate_up_proj_weight_quantizers.125 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.41e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.gate_up_proj_weight_quantizers.126 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.01e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.gate_up_proj_weight_quantizers.127 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.18e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.gate_up_proj_weight_quantizers.128 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.19e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.gate_up_proj_weight_quantizers.129 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.34e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.gate_up_proj_weight_quantizers.130 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.24e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.gate_up_proj_weight_quantizers.131 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.17e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.gate_up_proj_weight_quantizers.132 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.35e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.gate_up_proj_weight_quantizers.133 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.06e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.gate_up_proj_weight_quantizers.134 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.20e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.gate_up_proj_weight_quantizers.135 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.40e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.gate_up_proj_weight_quantizers.136 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.54e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.gate_up_proj_weight_quantizers.137 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.gate_up_proj_weight_quantizers.138 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.13e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.gate_up_proj_weight_quantizers.139 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.47e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.gate_up_proj_weight_quantizers.140 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.gate_up_proj_weight_quantizers.141 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.15e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.gate_up_proj_weight_quantizers.142 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.gate_up_proj_weight_quantizers.143 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.22e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.gate_up_proj_weight_quantizers.144 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.25e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.gate_up_proj_weight_quantizers.145 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.14e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.gate_up_proj_weight_quantizers.146 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.16e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.gate_up_proj_weight_quantizers.147 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.21e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.gate_up_proj_weight_quantizers.148 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.01e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.gate_up_proj_weight_quantizers.149 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.29e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.gate_up_proj_weight_quantizers.150 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.30e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.gate_up_proj_weight_quantizers.151 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.52e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.gate_up_proj_weight_quantizers.152 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.24e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.gate_up_proj_weight_quantizers.153 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.32e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.gate_up_proj_weight_quantizers.154 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.19e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.gate_up_proj_weight_quantizers.155 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.gate_up_proj_weight_quantizers.156 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.gate_up_proj_weight_quantizers.157 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.28e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.gate_up_proj_weight_quantizers.158 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.40e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.gate_up_proj_weight_quantizers.159 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.35e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.gate_up_proj_weight_quantizers.160 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.44e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.gate_up_proj_weight_quantizers.161 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.gate_up_proj_weight_quantizers.162 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.33e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.gate_up_proj_weight_quantizers.163 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.gate_up_proj_weight_quantizers.164 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.35e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.gate_up_proj_weight_quantizers.165 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.27e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.gate_up_proj_weight_quantizers.166 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.78e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.gate_up_proj_weight_quantizers.167 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.gate_up_proj_weight_quantizers.168 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.13e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.gate_up_proj_weight_quantizers.169 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.42e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.gate_up_proj_weight_quantizers.170 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.76e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.gate_up_proj_weight_quantizers.171 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.43e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.gate_up_proj_weight_quantizers.172 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.19e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.gate_up_proj_weight_quantizers.173 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.21e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.gate_up_proj_weight_quantizers.174 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.19e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.gate_up_proj_weight_quantizers.175 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.33e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.gate_up_proj_weight_quantizers.176 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.85e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.gate_up_proj_weight_quantizers.177 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.15e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.gate_up_proj_weight_quantizers.178 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.37e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.gate_up_proj_weight_quantizers.179 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.30e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.gate_up_proj_weight_quantizers.180 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.gate_up_proj_weight_quantizers.181 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.20e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.gate_up_proj_weight_quantizers.182 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.gate_up_proj_weight_quantizers.183 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.44e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.gate_up_proj_weight_quantizers.184 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.15e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.gate_up_proj_weight_quantizers.185 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.22e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.gate_up_proj_weight_quantizers.186 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.29e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.gate_up_proj_weight_quantizers.187 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.38e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.gate_up_proj_weight_quantizers.188 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.24e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.gate_up_proj_weight_quantizers.189 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.43e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.gate_up_proj_weight_quantizers.190 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.36e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.gate_up_proj_weight_quantizers.191 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.29e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.gate_up_proj_weight_quantizers.192 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.24e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.gate_up_proj_weight_quantizers.193 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.34e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.gate_up_proj_weight_quantizers.194 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.17e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.gate_up_proj_weight_quantizers.195 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.36e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.gate_up_proj_weight_quantizers.196 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.31e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.gate_up_proj_weight_quantizers.197 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.32e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.gate_up_proj_weight_quantizers.198 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.13e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.gate_up_proj_weight_quantizers.199 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.18e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.gate_up_proj_weight_quantizers.200 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.14e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.gate_up_proj_weight_quantizers.201 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.gate_up_proj_weight_quantizers.202 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.22e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.gate_up_proj_weight_quantizers.203 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.25e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.gate_up_proj_weight_quantizers.204 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.33e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.gate_up_proj_weight_quantizers.205 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.52e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.gate_up_proj_weight_quantizers.206 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.22e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.gate_up_proj_weight_quantizers.207 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.87e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.gate_up_proj_weight_quantizers.208 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.29e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.gate_up_proj_weight_quantizers.209 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.30e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.gate_up_proj_weight_quantizers.210 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.24e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.gate_up_proj_weight_quantizers.211 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.gate_up_proj_weight_quantizers.212 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.72e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.gate_up_proj_weight_quantizers.213 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.gate_up_proj_weight_quantizers.214 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.42e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.gate_up_proj_weight_quantizers.215 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.49e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.gate_up_proj_weight_quantizers.216 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.14e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.gate_up_proj_weight_quantizers.217 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.gate_up_proj_weight_quantizers.218 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.85e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.gate_up_proj_weight_quantizers.219 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.31e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.gate_up_proj_weight_quantizers.220 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.21e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.gate_up_proj_weight_quantizers.221 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.19e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.gate_up_proj_weight_quantizers.222 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.57e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.gate_up_proj_weight_quantizers.223 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.18e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.gate_up_proj_weight_quantizers.224 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.14e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.gate_up_proj_weight_quantizers.225 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.15e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.gate_up_proj_weight_quantizers.226 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.46e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.gate_up_proj_weight_quantizers.227 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.45e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.gate_up_proj_weight_quantizers.228 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.15e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.gate_up_proj_weight_quantizers.229 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.19e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.gate_up_proj_weight_quantizers.230 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.21e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.gate_up_proj_weight_quantizers.231 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.74e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.gate_up_proj_weight_quantizers.232 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.18e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.gate_up_proj_weight_quantizers.233 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.gate_up_proj_weight_quantizers.234 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.87e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.gate_up_proj_weight_quantizers.235 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.24e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.gate_up_proj_weight_quantizers.236 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.56e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.gate_up_proj_weight_quantizers.237 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.14e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.gate_up_proj_weight_quantizers.238 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.28e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.gate_up_proj_weight_quantizers.239 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.20e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.gate_up_proj_weight_quantizers.240 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.77e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.gate_up_proj_weight_quantizers.241 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.25e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.gate_up_proj_weight_quantizers.242 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.36e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.gate_up_proj_weight_quantizers.243 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.gate_up_proj_weight_quantizers.244 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.14e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.gate_up_proj_weight_quantizers.245 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.27e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.gate_up_proj_weight_quantizers.246 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.35e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.gate_up_proj_weight_quantizers.247 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.17e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.gate_up_proj_weight_quantizers.248 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.18e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.gate_up_proj_weight_quantizers.249 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.92e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.gate_up_proj_weight_quantizers.250 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.gate_up_proj_weight_quantizers.251 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.67e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.gate_up_proj_weight_quantizers.252 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.32e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.gate_up_proj_weight_quantizers.253 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.19e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.gate_up_proj_weight_quantizers.254 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.37e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.gate_up_proj_weight_quantizers.255 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.40e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.down_proj_input_quantizer TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.90e+00 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.down_proj_weight_quantizers.0 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.down_proj_weight_quantizers.1 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.24e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.down_proj_weight_quantizers.2 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.93e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.down_proj_weight_quantizers.3 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.down_proj_weight_quantizers.4 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.down_proj_weight_quantizers.5 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.down_proj_weight_quantizers.6 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.33e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.down_proj_weight_quantizers.7 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.down_proj_weight_quantizers.8 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.down_proj_weight_quantizers.9 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.52e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.down_proj_weight_quantizers.10 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.29e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.down_proj_weight_quantizers.11 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.78e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.down_proj_weight_quantizers.12 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.down_proj_weight_quantizers.13 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.29e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.down_proj_weight_quantizers.14 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.down_proj_weight_quantizers.15 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.50e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.down_proj_weight_quantizers.16 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.38e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.down_proj_weight_quantizers.17 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.down_proj_weight_quantizers.18 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.59e-02 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.down_proj_weight_quantizers.19 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.down_proj_weight_quantizers.20 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.74e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.down_proj_weight_quantizers.21 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.97e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.down_proj_weight_quantizers.22 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.15e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.down_proj_weight_quantizers.23 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.down_proj_weight_quantizers.24 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.down_proj_weight_quantizers.25 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.down_proj_weight_quantizers.26 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.21e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.down_proj_weight_quantizers.27 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.99e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.down_proj_weight_quantizers.28 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.down_proj_weight_quantizers.29 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.down_proj_weight_quantizers.30 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.down_proj_weight_quantizers.31 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.down_proj_weight_quantizers.32 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.32e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.down_proj_weight_quantizers.33 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.down_proj_weight_quantizers.34 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.38e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.down_proj_weight_quantizers.35 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.down_proj_weight_quantizers.36 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.51e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.down_proj_weight_quantizers.37 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.51e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.down_proj_weight_quantizers.38 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.down_proj_weight_quantizers.39 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.down_proj_weight_quantizers.40 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.down_proj_weight_quantizers.41 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.46e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.down_proj_weight_quantizers.42 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.49e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.down_proj_weight_quantizers.43 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.47e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.down_proj_weight_quantizers.44 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.45e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.down_proj_weight_quantizers.45 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.down_proj_weight_quantizers.46 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.20e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.down_proj_weight_quantizers.47 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.33e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.down_proj_weight_quantizers.48 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.down_proj_weight_quantizers.49 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.36e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.down_proj_weight_quantizers.50 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.14e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.down_proj_weight_quantizers.51 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.down_proj_weight_quantizers.52 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.down_proj_weight_quantizers.53 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.50e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.down_proj_weight_quantizers.54 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.22e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.down_proj_weight_quantizers.55 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.down_proj_weight_quantizers.56 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.57e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.down_proj_weight_quantizers.57 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.38e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.down_proj_weight_quantizers.58 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.down_proj_weight_quantizers.59 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.down_proj_weight_quantizers.60 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.down_proj_weight_quantizers.61 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.down_proj_weight_quantizers.62 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.44e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.down_proj_weight_quantizers.63 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.87e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.down_proj_weight_quantizers.64 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.30e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.down_proj_weight_quantizers.65 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.44e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.down_proj_weight_quantizers.66 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.50e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.down_proj_weight_quantizers.67 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.34e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.down_proj_weight_quantizers.68 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.down_proj_weight_quantizers.69 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.down_proj_weight_quantizers.70 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.37e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.down_proj_weight_quantizers.71 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.15e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.down_proj_weight_quantizers.72 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.down_proj_weight_quantizers.73 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.36e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.down_proj_weight_quantizers.74 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.24e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.down_proj_weight_quantizers.75 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.41e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.down_proj_weight_quantizers.76 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.14e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.down_proj_weight_quantizers.77 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.down_proj_weight_quantizers.78 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.47e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.down_proj_weight_quantizers.79 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.28e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.down_proj_weight_quantizers.80 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.down_proj_weight_quantizers.81 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.down_proj_weight_quantizers.82 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.56e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.down_proj_weight_quantizers.83 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.down_proj_weight_quantizers.84 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.61e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.down_proj_weight_quantizers.85 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.43e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.down_proj_weight_quantizers.86 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.down_proj_weight_quantizers.87 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.48e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.down_proj_weight_quantizers.88 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.down_proj_weight_quantizers.89 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.down_proj_weight_quantizers.90 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.27e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.down_proj_weight_quantizers.91 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.50e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.down_proj_weight_quantizers.92 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.22e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.down_proj_weight_quantizers.93 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.45e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.down_proj_weight_quantizers.94 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.down_proj_weight_quantizers.95 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.57e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.down_proj_weight_quantizers.96 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.down_proj_weight_quantizers.97 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.down_proj_weight_quantizers.98 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.down_proj_weight_quantizers.99 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.20e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.down_proj_weight_quantizers.100 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.down_proj_weight_quantizers.101 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.42e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.down_proj_weight_quantizers.102 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.64e-02 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.down_proj_weight_quantizers.103 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.41e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.down_proj_weight_quantizers.104 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.20e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.down_proj_weight_quantizers.105 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.down_proj_weight_quantizers.106 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.66e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.down_proj_weight_quantizers.107 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.down_proj_weight_quantizers.108 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.20e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.down_proj_weight_quantizers.109 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.down_proj_weight_quantizers.110 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.83e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.down_proj_weight_quantizers.111 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.74e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.down_proj_weight_quantizers.112 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.22e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.down_proj_weight_quantizers.113 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.21e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.down_proj_weight_quantizers.114 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.down_proj_weight_quantizers.115 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.22e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.down_proj_weight_quantizers.116 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.19e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.down_proj_weight_quantizers.117 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.25e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.down_proj_weight_quantizers.118 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.88e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.down_proj_weight_quantizers.119 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.87e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.down_proj_weight_quantizers.120 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=4.32e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.down_proj_weight_quantizers.121 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.31e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.down_proj_weight_quantizers.122 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.29e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.down_proj_weight_quantizers.123 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.down_proj_weight_quantizers.124 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.31e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.down_proj_weight_quantizers.125 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.16e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.down_proj_weight_quantizers.126 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.83e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.down_proj_weight_quantizers.127 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.21e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.down_proj_weight_quantizers.128 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.31e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.down_proj_weight_quantizers.129 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.45e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.down_proj_weight_quantizers.130 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.33e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.down_proj_weight_quantizers.131 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.down_proj_weight_quantizers.132 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.30e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.down_proj_weight_quantizers.133 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.28e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.down_proj_weight_quantizers.134 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.16e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.down_proj_weight_quantizers.135 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.down_proj_weight_quantizers.136 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.down_proj_weight_quantizers.137 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.16e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.down_proj_weight_quantizers.138 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.down_proj_weight_quantizers.139 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.47e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.down_proj_weight_quantizers.140 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.down_proj_weight_quantizers.141 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.27e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.down_proj_weight_quantizers.142 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.27e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.down_proj_weight_quantizers.143 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.43e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.down_proj_weight_quantizers.144 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.26e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.down_proj_weight_quantizers.145 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.43e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.down_proj_weight_quantizers.146 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.down_proj_weight_quantizers.147 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.22e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.down_proj_weight_quantizers.148 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.31e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.down_proj_weight_quantizers.149 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.20e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.down_proj_weight_quantizers.150 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.25e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.down_proj_weight_quantizers.151 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.43e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.down_proj_weight_quantizers.152 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.down_proj_weight_quantizers.153 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.43e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.down_proj_weight_quantizers.154 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.19e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.down_proj_weight_quantizers.155 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.17e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.down_proj_weight_quantizers.156 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.down_proj_weight_quantizers.157 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.37e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.down_proj_weight_quantizers.158 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.25e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.down_proj_weight_quantizers.159 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.down_proj_weight_quantizers.160 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.34e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.down_proj_weight_quantizers.161 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.72e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.down_proj_weight_quantizers.162 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.46e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.down_proj_weight_quantizers.163 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.down_proj_weight_quantizers.164 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.35e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.down_proj_weight_quantizers.165 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.71e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.down_proj_weight_quantizers.166 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.54e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.down_proj_weight_quantizers.167 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.60e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.down_proj_weight_quantizers.168 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.37e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.down_proj_weight_quantizers.169 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.40e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.down_proj_weight_quantizers.170 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.down_proj_weight_quantizers.171 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.21e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.down_proj_weight_quantizers.172 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.25e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.down_proj_weight_quantizers.173 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.00e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.down_proj_weight_quantizers.174 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.down_proj_weight_quantizers.175 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.20e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.down_proj_weight_quantizers.176 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.16e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.down_proj_weight_quantizers.177 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.46e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.down_proj_weight_quantizers.178 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.down_proj_weight_quantizers.179 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.down_proj_weight_quantizers.180 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.down_proj_weight_quantizers.181 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.down_proj_weight_quantizers.182 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.down_proj_weight_quantizers.183 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.60e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.down_proj_weight_quantizers.184 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.33e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.down_proj_weight_quantizers.185 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.20e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.down_proj_weight_quantizers.186 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.down_proj_weight_quantizers.187 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.21e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.down_proj_weight_quantizers.188 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.down_proj_weight_quantizers.189 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.27e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.down_proj_weight_quantizers.190 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.down_proj_weight_quantizers.191 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.95e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.down_proj_weight_quantizers.192 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.down_proj_weight_quantizers.193 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.39e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.down_proj_weight_quantizers.194 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.56e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.down_proj_weight_quantizers.195 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.88e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.down_proj_weight_quantizers.196 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.down_proj_weight_quantizers.197 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.82e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.down_proj_weight_quantizers.198 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.down_proj_weight_quantizers.199 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.24e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.down_proj_weight_quantizers.200 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.down_proj_weight_quantizers.201 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.down_proj_weight_quantizers.202 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.35e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.down_proj_weight_quantizers.203 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.down_proj_weight_quantizers.204 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.67e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.down_proj_weight_quantizers.205 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.82e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.down_proj_weight_quantizers.206 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.23e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.down_proj_weight_quantizers.207 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.66e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.down_proj_weight_quantizers.208 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.53e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.down_proj_weight_quantizers.209 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.24e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.down_proj_weight_quantizers.210 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.down_proj_weight_quantizers.211 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.67e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.down_proj_weight_quantizers.212 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.85e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.down_proj_weight_quantizers.213 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.down_proj_weight_quantizers.214 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.down_proj_weight_quantizers.215 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.38e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.down_proj_weight_quantizers.216 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.10e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.down_proj_weight_quantizers.217 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.16e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.down_proj_weight_quantizers.218 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.42e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.down_proj_weight_quantizers.219 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.47e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.down_proj_weight_quantizers.220 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.down_proj_weight_quantizers.221 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.down_proj_weight_quantizers.222 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.35e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.down_proj_weight_quantizers.223 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.down_proj_weight_quantizers.224 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.33e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.down_proj_weight_quantizers.225 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.down_proj_weight_quantizers.226 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.21e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.down_proj_weight_quantizers.227 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.28e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.down_proj_weight_quantizers.228 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.17e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.down_proj_weight_quantizers.229 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.33e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.down_proj_weight_quantizers.230 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.down_proj_weight_quantizers.231 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.25e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.down_proj_weight_quantizers.232 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.down_proj_weight_quantizers.233 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.down_proj_weight_quantizers.234 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.58e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.down_proj_weight_quantizers.235 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.down_proj_weight_quantizers.236 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.19e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.down_proj_weight_quantizers.237 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.down_proj_weight_quantizers.238 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.down_proj_weight_quantizers.239 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.down_proj_weight_quantizers.240 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.86e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.down_proj_weight_quantizers.241 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.51e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.down_proj_weight_quantizers.242 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.43e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.down_proj_weight_quantizers.243 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.down_proj_weight_quantizers.244 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.74e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.down_proj_weight_quantizers.245 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.down_proj_weight_quantizers.246 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.41e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.down_proj_weight_quantizers.247 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.16e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.down_proj_weight_quantizers.248 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.53e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.down_proj_weight_quantizers.249 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.87e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.down_proj_weight_quantizers.250 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.39e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.down_proj_weight_quantizers.251 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.41e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.down_proj_weight_quantizers.252 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.15e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.down_proj_weight_quantizers.253 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.33e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.down_proj_weight_quantizers.254 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.79e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.down_proj_weight_quantizers.255 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.60e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.shared_experts.gate_proj.input_quantizer TensorQuantizer(disabled) +model.layers.6.mlp.shared_experts.gate_proj.output_quantizer TensorQuantizer(disabled) +model.layers.6.mlp.shared_experts.gate_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.6.mlp.shared_experts.up_proj.input_quantizer TensorQuantizer(disabled) +model.layers.6.mlp.shared_experts.up_proj.output_quantizer TensorQuantizer(disabled) +model.layers.6.mlp.shared_experts.up_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.6.mlp.shared_experts.down_proj.input_quantizer TensorQuantizer(disabled) +model.layers.6.mlp.shared_experts.down_proj.output_quantizer TensorQuantizer(disabled) +model.layers.6.mlp.shared_experts.down_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.7.self_attn.q_a_proj.input_quantizer TensorQuantizer(disabled) +model.layers.7.self_attn.q_a_proj.output_quantizer TensorQuantizer(disabled) +model.layers.7.self_attn.q_a_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.7.self_attn.q_b_proj.input_quantizer TensorQuantizer(disabled) +model.layers.7.self_attn.q_b_proj.output_quantizer TensorQuantizer(disabled) +model.layers.7.self_attn.q_b_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.7.self_attn.kv_a_proj_with_mqa.input_quantizer TensorQuantizer(disabled) +model.layers.7.self_attn.kv_a_proj_with_mqa.output_quantizer TensorQuantizer(disabled) +model.layers.7.self_attn.kv_a_proj_with_mqa.weight_quantizer TensorQuantizer(disabled) +model.layers.7.self_attn.kv_b_proj.input_quantizer TensorQuantizer(disabled) +model.layers.7.self_attn.kv_b_proj.output_quantizer TensorQuantizer(disabled) +model.layers.7.self_attn.kv_b_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.7.self_attn.o_proj.input_quantizer TensorQuantizer(disabled) +model.layers.7.self_attn.o_proj.output_quantizer TensorQuantizer(disabled) +model.layers.7.self_attn.o_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.7.self_attn.q_bmm_quantizer TensorQuantizer(disabled) +model.layers.7.self_attn.k_bmm_quantizer TensorQuantizer((4, 3) bit fake per-tensor amax=4.48e+02(const) calibrator=MaxCalibrator quant) +model.layers.7.self_attn.v_bmm_quantizer TensorQuantizer((4, 3) bit fake per-tensor amax=4.48e+02(const) calibrator=MaxCalibrator quant) +model.layers.7.self_attn.p_bmm_quantizer TensorQuantizer(disabled) +model.layers.7.mlp.experts.gate_up_proj_input_quantizer TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.65e+00 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.gate_up_proj_weight_quantizers.0 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.gate_up_proj_weight_quantizers.1 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=4.14e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.gate_up_proj_weight_quantizers.2 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.gate_up_proj_weight_quantizers.3 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.gate_up_proj_weight_quantizers.4 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.97e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.gate_up_proj_weight_quantizers.5 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.gate_up_proj_weight_quantizers.6 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.gate_up_proj_weight_quantizers.7 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.15e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.gate_up_proj_weight_quantizers.8 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.24e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.gate_up_proj_weight_quantizers.9 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.gate_up_proj_weight_quantizers.10 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.gate_up_proj_weight_quantizers.11 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.gate_up_proj_weight_quantizers.12 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.gate_up_proj_weight_quantizers.13 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.gate_up_proj_weight_quantizers.14 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.18e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.gate_up_proj_weight_quantizers.15 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.gate_up_proj_weight_quantizers.16 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.14e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.gate_up_proj_weight_quantizers.17 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.gate_up_proj_weight_quantizers.18 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.gate_up_proj_weight_quantizers.19 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.gate_up_proj_weight_quantizers.20 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.24e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.gate_up_proj_weight_quantizers.21 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.gate_up_proj_weight_quantizers.22 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.gate_up_proj_weight_quantizers.23 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.gate_up_proj_weight_quantizers.24 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.gate_up_proj_weight_quantizers.25 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.gate_up_proj_weight_quantizers.26 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.gate_up_proj_weight_quantizers.27 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.77e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.gate_up_proj_weight_quantizers.28 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.gate_up_proj_weight_quantizers.29 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.gate_up_proj_weight_quantizers.30 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.gate_up_proj_weight_quantizers.31 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.gate_up_proj_weight_quantizers.32 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.gate_up_proj_weight_quantizers.33 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.gate_up_proj_weight_quantizers.34 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.gate_up_proj_weight_quantizers.35 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.gate_up_proj_weight_quantizers.36 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.14e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.gate_up_proj_weight_quantizers.37 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.gate_up_proj_weight_quantizers.38 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.13e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.gate_up_proj_weight_quantizers.39 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.gate_up_proj_weight_quantizers.40 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.16e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.gate_up_proj_weight_quantizers.41 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.37e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.gate_up_proj_weight_quantizers.42 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.gate_up_proj_weight_quantizers.43 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.gate_up_proj_weight_quantizers.44 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.gate_up_proj_weight_quantizers.45 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.gate_up_proj_weight_quantizers.46 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.gate_up_proj_weight_quantizers.47 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.24e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.gate_up_proj_weight_quantizers.48 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.gate_up_proj_weight_quantizers.49 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.gate_up_proj_weight_quantizers.50 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.gate_up_proj_weight_quantizers.51 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=6.68e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.gate_up_proj_weight_quantizers.52 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.gate_up_proj_weight_quantizers.53 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.gate_up_proj_weight_quantizers.54 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.gate_up_proj_weight_quantizers.55 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.gate_up_proj_weight_quantizers.56 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.18e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.gate_up_proj_weight_quantizers.57 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.gate_up_proj_weight_quantizers.58 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.gate_up_proj_weight_quantizers.59 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.gate_up_proj_weight_quantizers.60 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.13e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.gate_up_proj_weight_quantizers.61 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.gate_up_proj_weight_quantizers.62 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.32e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.gate_up_proj_weight_quantizers.63 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.gate_up_proj_weight_quantizers.64 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.gate_up_proj_weight_quantizers.65 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.gate_up_proj_weight_quantizers.66 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.55e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.gate_up_proj_weight_quantizers.67 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.gate_up_proj_weight_quantizers.68 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.gate_up_proj_weight_quantizers.69 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.gate_up_proj_weight_quantizers.70 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.gate_up_proj_weight_quantizers.71 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.gate_up_proj_weight_quantizers.72 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.gate_up_proj_weight_quantizers.73 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.gate_up_proj_weight_quantizers.74 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.gate_up_proj_weight_quantizers.75 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.gate_up_proj_weight_quantizers.76 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.gate_up_proj_weight_quantizers.77 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.gate_up_proj_weight_quantizers.78 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.39e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.gate_up_proj_weight_quantizers.79 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.gate_up_proj_weight_quantizers.80 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.59e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.gate_up_proj_weight_quantizers.81 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.gate_up_proj_weight_quantizers.82 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.gate_up_proj_weight_quantizers.83 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.gate_up_proj_weight_quantizers.84 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.20e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.gate_up_proj_weight_quantizers.85 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.gate_up_proj_weight_quantizers.86 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.gate_up_proj_weight_quantizers.87 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.gate_up_proj_weight_quantizers.88 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.21e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.gate_up_proj_weight_quantizers.89 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.gate_up_proj_weight_quantizers.90 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.gate_up_proj_weight_quantizers.91 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.gate_up_proj_weight_quantizers.92 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.33e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.gate_up_proj_weight_quantizers.93 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.gate_up_proj_weight_quantizers.94 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.gate_up_proj_weight_quantizers.95 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.gate_up_proj_weight_quantizers.96 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.gate_up_proj_weight_quantizers.97 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.gate_up_proj_weight_quantizers.98 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.gate_up_proj_weight_quantizers.99 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.gate_up_proj_weight_quantizers.100 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.14e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.gate_up_proj_weight_quantizers.101 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.16e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.gate_up_proj_weight_quantizers.102 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.gate_up_proj_weight_quantizers.103 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.gate_up_proj_weight_quantizers.104 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.gate_up_proj_weight_quantizers.105 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.gate_up_proj_weight_quantizers.106 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.gate_up_proj_weight_quantizers.107 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.gate_up_proj_weight_quantizers.108 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.gate_up_proj_weight_quantizers.109 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.gate_up_proj_weight_quantizers.110 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.13e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.gate_up_proj_weight_quantizers.111 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.gate_up_proj_weight_quantizers.112 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.gate_up_proj_weight_quantizers.113 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.gate_up_proj_weight_quantizers.114 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.19e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.gate_up_proj_weight_quantizers.115 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.17e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.gate_up_proj_weight_quantizers.116 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.gate_up_proj_weight_quantizers.117 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.gate_up_proj_weight_quantizers.118 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.gate_up_proj_weight_quantizers.119 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.gate_up_proj_weight_quantizers.120 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.gate_up_proj_weight_quantizers.121 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.gate_up_proj_weight_quantizers.122 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.15e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.gate_up_proj_weight_quantizers.123 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.51e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.gate_up_proj_weight_quantizers.124 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.17e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.gate_up_proj_weight_quantizers.125 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.gate_up_proj_weight_quantizers.126 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.gate_up_proj_weight_quantizers.127 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.gate_up_proj_weight_quantizers.128 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.gate_up_proj_weight_quantizers.129 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.gate_up_proj_weight_quantizers.130 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.15e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.gate_up_proj_weight_quantizers.131 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.gate_up_proj_weight_quantizers.132 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.gate_up_proj_weight_quantizers.133 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.gate_up_proj_weight_quantizers.134 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.gate_up_proj_weight_quantizers.135 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.gate_up_proj_weight_quantizers.136 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.19e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.gate_up_proj_weight_quantizers.137 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.26e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.gate_up_proj_weight_quantizers.138 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.gate_up_proj_weight_quantizers.139 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.gate_up_proj_weight_quantizers.140 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.gate_up_proj_weight_quantizers.141 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.gate_up_proj_weight_quantizers.142 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.21e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.gate_up_proj_weight_quantizers.143 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.gate_up_proj_weight_quantizers.144 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.gate_up_proj_weight_quantizers.145 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.gate_up_proj_weight_quantizers.146 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.gate_up_proj_weight_quantizers.147 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.gate_up_proj_weight_quantizers.148 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.gate_up_proj_weight_quantizers.149 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.gate_up_proj_weight_quantizers.150 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.gate_up_proj_weight_quantizers.151 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.gate_up_proj_weight_quantizers.152 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.gate_up_proj_weight_quantizers.153 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.gate_up_proj_weight_quantizers.154 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.48e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.gate_up_proj_weight_quantizers.155 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.30e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.gate_up_proj_weight_quantizers.156 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.gate_up_proj_weight_quantizers.157 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.gate_up_proj_weight_quantizers.158 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.gate_up_proj_weight_quantizers.159 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.14e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.gate_up_proj_weight_quantizers.160 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.gate_up_proj_weight_quantizers.161 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.gate_up_proj_weight_quantizers.162 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.gate_up_proj_weight_quantizers.163 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.gate_up_proj_weight_quantizers.164 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.gate_up_proj_weight_quantizers.165 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.gate_up_proj_weight_quantizers.166 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.gate_up_proj_weight_quantizers.167 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.gate_up_proj_weight_quantizers.168 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.gate_up_proj_weight_quantizers.169 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.gate_up_proj_weight_quantizers.170 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.gate_up_proj_weight_quantizers.171 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.22e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.gate_up_proj_weight_quantizers.172 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.gate_up_proj_weight_quantizers.173 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.gate_up_proj_weight_quantizers.174 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.gate_up_proj_weight_quantizers.175 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.gate_up_proj_weight_quantizers.176 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.38e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.gate_up_proj_weight_quantizers.177 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.gate_up_proj_weight_quantizers.178 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.16e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.gate_up_proj_weight_quantizers.179 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.22e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.gate_up_proj_weight_quantizers.180 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.gate_up_proj_weight_quantizers.181 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.gate_up_proj_weight_quantizers.182 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.38e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.gate_up_proj_weight_quantizers.183 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.gate_up_proj_weight_quantizers.184 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.gate_up_proj_weight_quantizers.185 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.gate_up_proj_weight_quantizers.186 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.gate_up_proj_weight_quantizers.187 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.19e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.gate_up_proj_weight_quantizers.188 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.33e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.gate_up_proj_weight_quantizers.189 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.gate_up_proj_weight_quantizers.190 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.gate_up_proj_weight_quantizers.191 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.gate_up_proj_weight_quantizers.192 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.gate_up_proj_weight_quantizers.193 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.gate_up_proj_weight_quantizers.194 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.gate_up_proj_weight_quantizers.195 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.gate_up_proj_weight_quantizers.196 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.gate_up_proj_weight_quantizers.197 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.gate_up_proj_weight_quantizers.198 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.52e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.gate_up_proj_weight_quantizers.199 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.gate_up_proj_weight_quantizers.200 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.gate_up_proj_weight_quantizers.201 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.gate_up_proj_weight_quantizers.202 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.14e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.gate_up_proj_weight_quantizers.203 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.gate_up_proj_weight_quantizers.204 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.gate_up_proj_weight_quantizers.205 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.gate_up_proj_weight_quantizers.206 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.gate_up_proj_weight_quantizers.207 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.gate_up_proj_weight_quantizers.208 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.13e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.gate_up_proj_weight_quantizers.209 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.gate_up_proj_weight_quantizers.210 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.14e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.gate_up_proj_weight_quantizers.211 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.gate_up_proj_weight_quantizers.212 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.gate_up_proj_weight_quantizers.213 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.19e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.gate_up_proj_weight_quantizers.214 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.14e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.gate_up_proj_weight_quantizers.215 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=7.70e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.gate_up_proj_weight_quantizers.216 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.gate_up_proj_weight_quantizers.217 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.gate_up_proj_weight_quantizers.218 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.14e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.gate_up_proj_weight_quantizers.219 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.gate_up_proj_weight_quantizers.220 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.gate_up_proj_weight_quantizers.221 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.gate_up_proj_weight_quantizers.222 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.gate_up_proj_weight_quantizers.223 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.gate_up_proj_weight_quantizers.224 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.gate_up_proj_weight_quantizers.225 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.25e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.gate_up_proj_weight_quantizers.226 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.gate_up_proj_weight_quantizers.227 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.gate_up_proj_weight_quantizers.228 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.gate_up_proj_weight_quantizers.229 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.gate_up_proj_weight_quantizers.230 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.gate_up_proj_weight_quantizers.231 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.gate_up_proj_weight_quantizers.232 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.gate_up_proj_weight_quantizers.233 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.gate_up_proj_weight_quantizers.234 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.gate_up_proj_weight_quantizers.235 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.gate_up_proj_weight_quantizers.236 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.13e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.gate_up_proj_weight_quantizers.237 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.gate_up_proj_weight_quantizers.238 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.gate_up_proj_weight_quantizers.239 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.gate_up_proj_weight_quantizers.240 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.gate_up_proj_weight_quantizers.241 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.gate_up_proj_weight_quantizers.242 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.gate_up_proj_weight_quantizers.243 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.gate_up_proj_weight_quantizers.244 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.24e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.gate_up_proj_weight_quantizers.245 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.gate_up_proj_weight_quantizers.246 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.17e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.gate_up_proj_weight_quantizers.247 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.gate_up_proj_weight_quantizers.248 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.gate_up_proj_weight_quantizers.249 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.gate_up_proj_weight_quantizers.250 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.30e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.gate_up_proj_weight_quantizers.251 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.gate_up_proj_weight_quantizers.252 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.gate_up_proj_weight_quantizers.253 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.gate_up_proj_weight_quantizers.254 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.14e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.gate_up_proj_weight_quantizers.255 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.down_proj_input_quantizer TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.20e+00 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.down_proj_weight_quantizers.0 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.46e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.down_proj_weight_quantizers.1 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.24e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.down_proj_weight_quantizers.2 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.down_proj_weight_quantizers.3 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.14e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.down_proj_weight_quantizers.4 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.89e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.down_proj_weight_quantizers.5 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.79e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.down_proj_weight_quantizers.6 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.68e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.down_proj_weight_quantizers.7 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.64e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.down_proj_weight_quantizers.8 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.down_proj_weight_quantizers.9 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.down_proj_weight_quantizers.10 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.57e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.down_proj_weight_quantizers.11 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.down_proj_weight_quantizers.12 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.down_proj_weight_quantizers.13 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.53e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.down_proj_weight_quantizers.14 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.03e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.down_proj_weight_quantizers.15 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.down_proj_weight_quantizers.16 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.41e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.down_proj_weight_quantizers.17 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.down_proj_weight_quantizers.18 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.down_proj_weight_quantizers.19 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.down_proj_weight_quantizers.20 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.32e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.down_proj_weight_quantizers.21 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.96e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.down_proj_weight_quantizers.22 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.21e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.down_proj_weight_quantizers.23 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.26e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.down_proj_weight_quantizers.24 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.down_proj_weight_quantizers.25 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.down_proj_weight_quantizers.26 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.16e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.down_proj_weight_quantizers.27 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.60e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.down_proj_weight_quantizers.28 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.down_proj_weight_quantizers.29 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.21e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.down_proj_weight_quantizers.30 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.down_proj_weight_quantizers.31 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.down_proj_weight_quantizers.32 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.down_proj_weight_quantizers.33 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.down_proj_weight_quantizers.34 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.36e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.down_proj_weight_quantizers.35 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.down_proj_weight_quantizers.36 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.43e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.down_proj_weight_quantizers.37 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.down_proj_weight_quantizers.38 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.down_proj_weight_quantizers.39 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.96e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.down_proj_weight_quantizers.40 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.27e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.down_proj_weight_quantizers.41 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.53e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.down_proj_weight_quantizers.42 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.down_proj_weight_quantizers.43 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.down_proj_weight_quantizers.44 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.down_proj_weight_quantizers.45 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.20e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.down_proj_weight_quantizers.46 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.down_proj_weight_quantizers.47 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.22e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.down_proj_weight_quantizers.48 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.down_proj_weight_quantizers.49 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.14e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.down_proj_weight_quantizers.50 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.down_proj_weight_quantizers.51 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e+00 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.down_proj_weight_quantizers.52 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.down_proj_weight_quantizers.53 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.17e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.down_proj_weight_quantizers.54 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.down_proj_weight_quantizers.55 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.14e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.down_proj_weight_quantizers.56 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.down_proj_weight_quantizers.57 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.18e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.down_proj_weight_quantizers.58 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.down_proj_weight_quantizers.59 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.49e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.down_proj_weight_quantizers.60 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.51e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.down_proj_weight_quantizers.61 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.down_proj_weight_quantizers.62 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.01e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.down_proj_weight_quantizers.63 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.24e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.down_proj_weight_quantizers.64 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.down_proj_weight_quantizers.65 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.73e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.down_proj_weight_quantizers.66 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.78e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.down_proj_weight_quantizers.67 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.down_proj_weight_quantizers.68 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.down_proj_weight_quantizers.69 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.87e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.down_proj_weight_quantizers.70 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.down_proj_weight_quantizers.71 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.down_proj_weight_quantizers.72 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.down_proj_weight_quantizers.73 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.46e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.down_proj_weight_quantizers.74 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.down_proj_weight_quantizers.75 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.down_proj_weight_quantizers.76 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.down_proj_weight_quantizers.77 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.23e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.down_proj_weight_quantizers.78 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=4.79e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.down_proj_weight_quantizers.79 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.down_proj_weight_quantizers.80 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.33e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.down_proj_weight_quantizers.81 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.down_proj_weight_quantizers.82 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.down_proj_weight_quantizers.83 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.down_proj_weight_quantizers.84 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.down_proj_weight_quantizers.85 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.34e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.down_proj_weight_quantizers.86 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.down_proj_weight_quantizers.87 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.22e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.down_proj_weight_quantizers.88 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.64e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.down_proj_weight_quantizers.89 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.down_proj_weight_quantizers.90 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.down_proj_weight_quantizers.91 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.17e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.down_proj_weight_quantizers.92 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.34e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.down_proj_weight_quantizers.93 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.28e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.down_proj_weight_quantizers.94 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.down_proj_weight_quantizers.95 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.down_proj_weight_quantizers.96 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.73e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.down_proj_weight_quantizers.97 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.down_proj_weight_quantizers.98 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.down_proj_weight_quantizers.99 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.down_proj_weight_quantizers.100 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.28e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.down_proj_weight_quantizers.101 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.34e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.down_proj_weight_quantizers.102 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.down_proj_weight_quantizers.103 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.25e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.down_proj_weight_quantizers.104 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.down_proj_weight_quantizers.105 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.61e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.down_proj_weight_quantizers.106 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.down_proj_weight_quantizers.107 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.down_proj_weight_quantizers.108 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.31e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.down_proj_weight_quantizers.109 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.down_proj_weight_quantizers.110 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.40e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.down_proj_weight_quantizers.111 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.down_proj_weight_quantizers.112 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=5.94e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.down_proj_weight_quantizers.113 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.down_proj_weight_quantizers.114 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.20e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.down_proj_weight_quantizers.115 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.down_proj_weight_quantizers.116 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.down_proj_weight_quantizers.117 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.07e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.down_proj_weight_quantizers.118 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.down_proj_weight_quantizers.119 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.down_proj_weight_quantizers.120 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.down_proj_weight_quantizers.121 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.down_proj_weight_quantizers.122 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.down_proj_weight_quantizers.123 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.03e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.down_proj_weight_quantizers.124 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.54e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.down_proj_weight_quantizers.125 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.down_proj_weight_quantizers.126 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.down_proj_weight_quantizers.127 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.24e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.down_proj_weight_quantizers.128 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.59e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.down_proj_weight_quantizers.129 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.down_proj_weight_quantizers.130 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.28e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.down_proj_weight_quantizers.131 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.down_proj_weight_quantizers.132 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.down_proj_weight_quantizers.133 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.down_proj_weight_quantizers.134 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.down_proj_weight_quantizers.135 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.down_proj_weight_quantizers.136 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.down_proj_weight_quantizers.137 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.40e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.down_proj_weight_quantizers.138 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.47e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.down_proj_weight_quantizers.139 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.down_proj_weight_quantizers.140 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.down_proj_weight_quantizers.141 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.18e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.down_proj_weight_quantizers.142 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.55e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.down_proj_weight_quantizers.143 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.down_proj_weight_quantizers.144 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.down_proj_weight_quantizers.145 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.down_proj_weight_quantizers.146 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.64e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.down_proj_weight_quantizers.147 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.down_proj_weight_quantizers.148 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.down_proj_weight_quantizers.149 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.28e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.down_proj_weight_quantizers.150 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.16e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.down_proj_weight_quantizers.151 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.down_proj_weight_quantizers.152 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.down_proj_weight_quantizers.153 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.down_proj_weight_quantizers.154 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.48e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.down_proj_weight_quantizers.155 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.95e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.down_proj_weight_quantizers.156 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.down_proj_weight_quantizers.157 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.down_proj_weight_quantizers.158 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.down_proj_weight_quantizers.159 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.down_proj_weight_quantizers.160 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.down_proj_weight_quantizers.161 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.down_proj_weight_quantizers.162 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.down_proj_weight_quantizers.163 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.down_proj_weight_quantizers.164 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.down_proj_weight_quantizers.165 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.down_proj_weight_quantizers.166 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.31e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.down_proj_weight_quantizers.167 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.down_proj_weight_quantizers.168 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.down_proj_weight_quantizers.169 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.66e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.down_proj_weight_quantizers.170 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.down_proj_weight_quantizers.171 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.15e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.down_proj_weight_quantizers.172 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.43e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.down_proj_weight_quantizers.173 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.down_proj_weight_quantizers.174 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.down_proj_weight_quantizers.175 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.down_proj_weight_quantizers.176 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.down_proj_weight_quantizers.177 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.56e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.down_proj_weight_quantizers.178 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.down_proj_weight_quantizers.179 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.22e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.down_proj_weight_quantizers.180 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.49e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.down_proj_weight_quantizers.181 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.down_proj_weight_quantizers.182 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.19e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.down_proj_weight_quantizers.183 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.down_proj_weight_quantizers.184 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.down_proj_weight_quantizers.185 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.78e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.down_proj_weight_quantizers.186 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.44e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.down_proj_weight_quantizers.187 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.42e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.down_proj_weight_quantizers.188 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.24e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.down_proj_weight_quantizers.189 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.down_proj_weight_quantizers.190 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.down_proj_weight_quantizers.191 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.down_proj_weight_quantizers.192 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.down_proj_weight_quantizers.193 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.down_proj_weight_quantizers.194 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.down_proj_weight_quantizers.195 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.down_proj_weight_quantizers.196 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.down_proj_weight_quantizers.197 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.down_proj_weight_quantizers.198 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=4.75e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.down_proj_weight_quantizers.199 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.28e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.down_proj_weight_quantizers.200 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.down_proj_weight_quantizers.201 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.down_proj_weight_quantizers.202 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.93e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.down_proj_weight_quantizers.203 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.down_proj_weight_quantizers.204 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.down_proj_weight_quantizers.205 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.down_proj_weight_quantizers.206 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.43e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.down_proj_weight_quantizers.207 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.down_proj_weight_quantizers.208 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.75e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.down_proj_weight_quantizers.209 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.down_proj_weight_quantizers.210 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.down_proj_weight_quantizers.211 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.down_proj_weight_quantizers.212 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.77e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.down_proj_weight_quantizers.213 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.down_proj_weight_quantizers.214 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.45e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.down_proj_weight_quantizers.215 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=6.29e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.down_proj_weight_quantizers.216 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.down_proj_weight_quantizers.217 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.down_proj_weight_quantizers.218 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.38e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.down_proj_weight_quantizers.219 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.down_proj_weight_quantizers.220 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.down_proj_weight_quantizers.221 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.down_proj_weight_quantizers.222 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.down_proj_weight_quantizers.223 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.49e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.down_proj_weight_quantizers.224 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.76e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.down_proj_weight_quantizers.225 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.46e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.down_proj_weight_quantizers.226 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.down_proj_weight_quantizers.227 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.down_proj_weight_quantizers.228 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.down_proj_weight_quantizers.229 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.22e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.down_proj_weight_quantizers.230 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.down_proj_weight_quantizers.231 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.down_proj_weight_quantizers.232 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.down_proj_weight_quantizers.233 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.down_proj_weight_quantizers.234 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.down_proj_weight_quantizers.235 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.down_proj_weight_quantizers.236 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.20e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.down_proj_weight_quantizers.237 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.28e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.down_proj_weight_quantizers.238 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.down_proj_weight_quantizers.239 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.down_proj_weight_quantizers.240 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.down_proj_weight_quantizers.241 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.down_proj_weight_quantizers.242 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.down_proj_weight_quantizers.243 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.13e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.down_proj_weight_quantizers.244 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.34e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.down_proj_weight_quantizers.245 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.23e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.down_proj_weight_quantizers.246 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.32e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.down_proj_weight_quantizers.247 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.down_proj_weight_quantizers.248 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.17e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.down_proj_weight_quantizers.249 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.29e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.down_proj_weight_quantizers.250 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.27e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.down_proj_weight_quantizers.251 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.down_proj_weight_quantizers.252 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.down_proj_weight_quantizers.253 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.60e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.down_proj_weight_quantizers.254 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.80e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.down_proj_weight_quantizers.255 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.shared_experts.gate_proj.input_quantizer TensorQuantizer(disabled) +model.layers.7.mlp.shared_experts.gate_proj.output_quantizer TensorQuantizer(disabled) +model.layers.7.mlp.shared_experts.gate_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.7.mlp.shared_experts.up_proj.input_quantizer TensorQuantizer(disabled) +model.layers.7.mlp.shared_experts.up_proj.output_quantizer TensorQuantizer(disabled) +model.layers.7.mlp.shared_experts.up_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.7.mlp.shared_experts.down_proj.input_quantizer TensorQuantizer(disabled) +model.layers.7.mlp.shared_experts.down_proj.output_quantizer TensorQuantizer(disabled) +model.layers.7.mlp.shared_experts.down_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.8.self_attn.q_a_proj.input_quantizer TensorQuantizer(disabled) +model.layers.8.self_attn.q_a_proj.output_quantizer TensorQuantizer(disabled) +model.layers.8.self_attn.q_a_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.8.self_attn.q_b_proj.input_quantizer TensorQuantizer(disabled) +model.layers.8.self_attn.q_b_proj.output_quantizer TensorQuantizer(disabled) +model.layers.8.self_attn.q_b_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.8.self_attn.kv_a_proj_with_mqa.input_quantizer TensorQuantizer(disabled) +model.layers.8.self_attn.kv_a_proj_with_mqa.output_quantizer TensorQuantizer(disabled) +model.layers.8.self_attn.kv_a_proj_with_mqa.weight_quantizer TensorQuantizer(disabled) +model.layers.8.self_attn.kv_b_proj.input_quantizer TensorQuantizer(disabled) +model.layers.8.self_attn.kv_b_proj.output_quantizer TensorQuantizer(disabled) +model.layers.8.self_attn.kv_b_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.8.self_attn.o_proj.input_quantizer TensorQuantizer(disabled) +model.layers.8.self_attn.o_proj.output_quantizer TensorQuantizer(disabled) +model.layers.8.self_attn.o_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.8.self_attn.q_bmm_quantizer TensorQuantizer(disabled) +model.layers.8.self_attn.k_bmm_quantizer TensorQuantizer((4, 3) bit fake per-tensor amax=4.48e+02(const) calibrator=MaxCalibrator quant) +model.layers.8.self_attn.v_bmm_quantizer TensorQuantizer((4, 3) bit fake per-tensor amax=4.48e+02(const) calibrator=MaxCalibrator quant) +model.layers.8.self_attn.p_bmm_quantizer TensorQuantizer(disabled) +model.layers.8.mlp.experts.gate_up_proj_input_quantizer TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e+00 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.gate_up_proj_weight_quantizers.0 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.68e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.gate_up_proj_weight_quantizers.1 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.16e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.gate_up_proj_weight_quantizers.2 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.76e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.gate_up_proj_weight_quantizers.3 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.gate_up_proj_weight_quantizers.4 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.95e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.gate_up_proj_weight_quantizers.5 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.24e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.gate_up_proj_weight_quantizers.6 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.gate_up_proj_weight_quantizers.7 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.gate_up_proj_weight_quantizers.8 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.gate_up_proj_weight_quantizers.9 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.02e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.gate_up_proj_weight_quantizers.10 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.34e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.gate_up_proj_weight_quantizers.11 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.gate_up_proj_weight_quantizers.12 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.17e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.gate_up_proj_weight_quantizers.13 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.03e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.gate_up_proj_weight_quantizers.14 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.gate_up_proj_weight_quantizers.15 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.75e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.gate_up_proj_weight_quantizers.16 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.gate_up_proj_weight_quantizers.17 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.24e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.gate_up_proj_weight_quantizers.18 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.91e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.gate_up_proj_weight_quantizers.19 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.97e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.gate_up_proj_weight_quantizers.20 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.50e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.gate_up_proj_weight_quantizers.21 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.82e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.gate_up_proj_weight_quantizers.22 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.25e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.gate_up_proj_weight_quantizers.23 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.gate_up_proj_weight_quantizers.24 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.23e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.gate_up_proj_weight_quantizers.25 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.gate_up_proj_weight_quantizers.26 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.gate_up_proj_weight_quantizers.27 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.93e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.gate_up_proj_weight_quantizers.28 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.gate_up_proj_weight_quantizers.29 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.34e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.gate_up_proj_weight_quantizers.30 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.gate_up_proj_weight_quantizers.31 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.gate_up_proj_weight_quantizers.32 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.gate_up_proj_weight_quantizers.33 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.83e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.gate_up_proj_weight_quantizers.34 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.14e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.gate_up_proj_weight_quantizers.35 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.gate_up_proj_weight_quantizers.36 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.gate_up_proj_weight_quantizers.37 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.gate_up_proj_weight_quantizers.38 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.gate_up_proj_weight_quantizers.39 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.44e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.gate_up_proj_weight_quantizers.40 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.66e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.gate_up_proj_weight_quantizers.41 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.14e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.gate_up_proj_weight_quantizers.42 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.gate_up_proj_weight_quantizers.43 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.23e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.gate_up_proj_weight_quantizers.44 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.61e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.gate_up_proj_weight_quantizers.45 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.29e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.gate_up_proj_weight_quantizers.46 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.gate_up_proj_weight_quantizers.47 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.gate_up_proj_weight_quantizers.48 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.20e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.gate_up_proj_weight_quantizers.49 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.95e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.gate_up_proj_weight_quantizers.50 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.gate_up_proj_weight_quantizers.51 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.29e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.gate_up_proj_weight_quantizers.52 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.gate_up_proj_weight_quantizers.53 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.78e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.gate_up_proj_weight_quantizers.54 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.gate_up_proj_weight_quantizers.55 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.24e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.gate_up_proj_weight_quantizers.56 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.19e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.gate_up_proj_weight_quantizers.57 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.54e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.gate_up_proj_weight_quantizers.58 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.64e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.gate_up_proj_weight_quantizers.59 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.gate_up_proj_weight_quantizers.60 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.gate_up_proj_weight_quantizers.61 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.64e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.gate_up_proj_weight_quantizers.62 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.gate_up_proj_weight_quantizers.63 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.gate_up_proj_weight_quantizers.64 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.10e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.gate_up_proj_weight_quantizers.65 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.gate_up_proj_weight_quantizers.66 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.14e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.gate_up_proj_weight_quantizers.67 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.25e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.gate_up_proj_weight_quantizers.68 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.64e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.gate_up_proj_weight_quantizers.69 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.58e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.gate_up_proj_weight_quantizers.70 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.gate_up_proj_weight_quantizers.71 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.gate_up_proj_weight_quantizers.72 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.03e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.gate_up_proj_weight_quantizers.73 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.gate_up_proj_weight_quantizers.74 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.07e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.gate_up_proj_weight_quantizers.75 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.gate_up_proj_weight_quantizers.76 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.23e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.gate_up_proj_weight_quantizers.77 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.gate_up_proj_weight_quantizers.78 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.13e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.gate_up_proj_weight_quantizers.79 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.74e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.gate_up_proj_weight_quantizers.80 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.gate_up_proj_weight_quantizers.81 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.72e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.gate_up_proj_weight_quantizers.82 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.20e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.gate_up_proj_weight_quantizers.83 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.gate_up_proj_weight_quantizers.84 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.36e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.gate_up_proj_weight_quantizers.85 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.gate_up_proj_weight_quantizers.86 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.gate_up_proj_weight_quantizers.87 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.gate_up_proj_weight_quantizers.88 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.62e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.gate_up_proj_weight_quantizers.89 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.gate_up_proj_weight_quantizers.90 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.gate_up_proj_weight_quantizers.91 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.20e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.gate_up_proj_weight_quantizers.92 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.gate_up_proj_weight_quantizers.93 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.24e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.gate_up_proj_weight_quantizers.94 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.gate_up_proj_weight_quantizers.95 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.gate_up_proj_weight_quantizers.96 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.39e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.gate_up_proj_weight_quantizers.97 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.81e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.gate_up_proj_weight_quantizers.98 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.66e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.gate_up_proj_weight_quantizers.99 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.32e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.gate_up_proj_weight_quantizers.100 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.gate_up_proj_weight_quantizers.101 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.96e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.gate_up_proj_weight_quantizers.102 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=6.17e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.gate_up_proj_weight_quantizers.103 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.gate_up_proj_weight_quantizers.104 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.96e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.gate_up_proj_weight_quantizers.105 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.gate_up_proj_weight_quantizers.106 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.27e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.gate_up_proj_weight_quantizers.107 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.gate_up_proj_weight_quantizers.108 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.gate_up_proj_weight_quantizers.109 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.25e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.gate_up_proj_weight_quantizers.110 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.gate_up_proj_weight_quantizers.111 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.gate_up_proj_weight_quantizers.112 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.gate_up_proj_weight_quantizers.113 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.23e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.gate_up_proj_weight_quantizers.114 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.gate_up_proj_weight_quantizers.115 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.gate_up_proj_weight_quantizers.116 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.28e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.gate_up_proj_weight_quantizers.117 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.73e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.gate_up_proj_weight_quantizers.118 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.gate_up_proj_weight_quantizers.119 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.gate_up_proj_weight_quantizers.120 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.gate_up_proj_weight_quantizers.121 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.gate_up_proj_weight_quantizers.122 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.86e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.gate_up_proj_weight_quantizers.123 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.gate_up_proj_weight_quantizers.124 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.60e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.gate_up_proj_weight_quantizers.125 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.52e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.gate_up_proj_weight_quantizers.126 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.14e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.gate_up_proj_weight_quantizers.127 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.69e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.gate_up_proj_weight_quantizers.128 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.43e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.gate_up_proj_weight_quantizers.129 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.gate_up_proj_weight_quantizers.130 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.22e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.gate_up_proj_weight_quantizers.131 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.13e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.gate_up_proj_weight_quantizers.132 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.53e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.gate_up_proj_weight_quantizers.133 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.gate_up_proj_weight_quantizers.134 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.gate_up_proj_weight_quantizers.135 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.74e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.gate_up_proj_weight_quantizers.136 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.11e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.gate_up_proj_weight_quantizers.137 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.61e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.gate_up_proj_weight_quantizers.138 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.gate_up_proj_weight_quantizers.139 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.gate_up_proj_weight_quantizers.140 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=6.80e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.gate_up_proj_weight_quantizers.141 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.gate_up_proj_weight_quantizers.142 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.gate_up_proj_weight_quantizers.143 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.43e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.gate_up_proj_weight_quantizers.144 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.gate_up_proj_weight_quantizers.145 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.39e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.gate_up_proj_weight_quantizers.146 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.gate_up_proj_weight_quantizers.147 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.45e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.gate_up_proj_weight_quantizers.148 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.20e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.gate_up_proj_weight_quantizers.149 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.gate_up_proj_weight_quantizers.150 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.gate_up_proj_weight_quantizers.151 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.22e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.gate_up_proj_weight_quantizers.152 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.04e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.gate_up_proj_weight_quantizers.153 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.gate_up_proj_weight_quantizers.154 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.42e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.gate_up_proj_weight_quantizers.155 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.20e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.gate_up_proj_weight_quantizers.156 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.gate_up_proj_weight_quantizers.157 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.gate_up_proj_weight_quantizers.158 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.25e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.gate_up_proj_weight_quantizers.159 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.73e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.gate_up_proj_weight_quantizers.160 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.gate_up_proj_weight_quantizers.161 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.18e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.gate_up_proj_weight_quantizers.162 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.gate_up_proj_weight_quantizers.163 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.gate_up_proj_weight_quantizers.164 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.gate_up_proj_weight_quantizers.165 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.80e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.gate_up_proj_weight_quantizers.166 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.gate_up_proj_weight_quantizers.167 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.gate_up_proj_weight_quantizers.168 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.52e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.gate_up_proj_weight_quantizers.169 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.gate_up_proj_weight_quantizers.170 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.34e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.gate_up_proj_weight_quantizers.171 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.21e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.gate_up_proj_weight_quantizers.172 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.42e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.gate_up_proj_weight_quantizers.173 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.20e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.gate_up_proj_weight_quantizers.174 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.gate_up_proj_weight_quantizers.175 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.16e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.gate_up_proj_weight_quantizers.176 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.gate_up_proj_weight_quantizers.177 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.25e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.gate_up_proj_weight_quantizers.178 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.16e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.gate_up_proj_weight_quantizers.179 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.gate_up_proj_weight_quantizers.180 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.gate_up_proj_weight_quantizers.181 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.05e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.gate_up_proj_weight_quantizers.182 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.12e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.gate_up_proj_weight_quantizers.183 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.gate_up_proj_weight_quantizers.184 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.60e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.gate_up_proj_weight_quantizers.185 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.gate_up_proj_weight_quantizers.186 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.16e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.gate_up_proj_weight_quantizers.187 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.gate_up_proj_weight_quantizers.188 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.95e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.gate_up_proj_weight_quantizers.189 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.gate_up_proj_weight_quantizers.190 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.gate_up_proj_weight_quantizers.191 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.46e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.gate_up_proj_weight_quantizers.192 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.gate_up_proj_weight_quantizers.193 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.23e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.gate_up_proj_weight_quantizers.194 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.88e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.gate_up_proj_weight_quantizers.195 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.85e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.gate_up_proj_weight_quantizers.196 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.gate_up_proj_weight_quantizers.197 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.gate_up_proj_weight_quantizers.198 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.gate_up_proj_weight_quantizers.199 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.43e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.gate_up_proj_weight_quantizers.200 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.29e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.gate_up_proj_weight_quantizers.201 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.18e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.gate_up_proj_weight_quantizers.202 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.gate_up_proj_weight_quantizers.203 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.gate_up_proj_weight_quantizers.204 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.gate_up_proj_weight_quantizers.205 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.79e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.gate_up_proj_weight_quantizers.206 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=6.76e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.gate_up_proj_weight_quantizers.207 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.82e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.gate_up_proj_weight_quantizers.208 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.18e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.gate_up_proj_weight_quantizers.209 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.08e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.gate_up_proj_weight_quantizers.210 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.gate_up_proj_weight_quantizers.211 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=4.20e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.gate_up_proj_weight_quantizers.212 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.gate_up_proj_weight_quantizers.213 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.gate_up_proj_weight_quantizers.214 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.gate_up_proj_weight_quantizers.215 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.89e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.gate_up_proj_weight_quantizers.216 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.77e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.gate_up_proj_weight_quantizers.217 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.93e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.gate_up_proj_weight_quantizers.218 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.gate_up_proj_weight_quantizers.219 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.07e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.gate_up_proj_weight_quantizers.220 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.12e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.gate_up_proj_weight_quantizers.221 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.gate_up_proj_weight_quantizers.222 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.gate_up_proj_weight_quantizers.223 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.17e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.gate_up_proj_weight_quantizers.224 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.gate_up_proj_weight_quantizers.225 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.31e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.gate_up_proj_weight_quantizers.226 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.62e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.gate_up_proj_weight_quantizers.227 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.92e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.gate_up_proj_weight_quantizers.228 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.gate_up_proj_weight_quantizers.229 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.01e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.gate_up_proj_weight_quantizers.230 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.34e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.gate_up_proj_weight_quantizers.231 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.gate_up_proj_weight_quantizers.232 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.gate_up_proj_weight_quantizers.233 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.gate_up_proj_weight_quantizers.234 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.26e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.gate_up_proj_weight_quantizers.235 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.gate_up_proj_weight_quantizers.236 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.gate_up_proj_weight_quantizers.237 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.76e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.gate_up_proj_weight_quantizers.238 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.gate_up_proj_weight_quantizers.239 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.37e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.gate_up_proj_weight_quantizers.240 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.gate_up_proj_weight_quantizers.241 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.18e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.gate_up_proj_weight_quantizers.242 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.15e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.gate_up_proj_weight_quantizers.243 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.25e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.gate_up_proj_weight_quantizers.244 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.gate_up_proj_weight_quantizers.245 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.gate_up_proj_weight_quantizers.246 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.gate_up_proj_weight_quantizers.247 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.gate_up_proj_weight_quantizers.248 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.gate_up_proj_weight_quantizers.249 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.gate_up_proj_weight_quantizers.250 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.17e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.gate_up_proj_weight_quantizers.251 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.gate_up_proj_weight_quantizers.252 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.gate_up_proj_weight_quantizers.253 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.gate_up_proj_weight_quantizers.254 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.68e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.gate_up_proj_weight_quantizers.255 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.down_proj_input_quantizer TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e+02 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.down_proj_weight_quantizers.0 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.down_proj_weight_quantizers.1 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.down_proj_weight_quantizers.2 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.down_proj_weight_quantizers.3 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.down_proj_weight_quantizers.4 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.70e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.down_proj_weight_quantizers.5 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.22e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.down_proj_weight_quantizers.6 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.down_proj_weight_quantizers.7 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.down_proj_weight_quantizers.8 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.down_proj_weight_quantizers.9 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.26e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.down_proj_weight_quantizers.10 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.down_proj_weight_quantizers.11 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.down_proj_weight_quantizers.12 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.down_proj_weight_quantizers.13 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.43e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.down_proj_weight_quantizers.14 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.23e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.down_proj_weight_quantizers.15 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.35e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.down_proj_weight_quantizers.16 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.46e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.down_proj_weight_quantizers.17 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.down_proj_weight_quantizers.18 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.down_proj_weight_quantizers.19 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.down_proj_weight_quantizers.20 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.34e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.down_proj_weight_quantizers.21 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.down_proj_weight_quantizers.22 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.28e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.down_proj_weight_quantizers.23 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.down_proj_weight_quantizers.24 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.down_proj_weight_quantizers.25 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.52e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.down_proj_weight_quantizers.26 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.down_proj_weight_quantizers.27 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.79e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.down_proj_weight_quantizers.28 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.down_proj_weight_quantizers.29 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.down_proj_weight_quantizers.30 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.down_proj_weight_quantizers.31 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.68e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.down_proj_weight_quantizers.32 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.down_proj_weight_quantizers.33 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.69e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.down_proj_weight_quantizers.34 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.down_proj_weight_quantizers.35 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.35e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.down_proj_weight_quantizers.36 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.down_proj_weight_quantizers.37 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.14e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.down_proj_weight_quantizers.38 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.down_proj_weight_quantizers.39 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.down_proj_weight_quantizers.40 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.down_proj_weight_quantizers.41 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.down_proj_weight_quantizers.42 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.down_proj_weight_quantizers.43 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.43e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.down_proj_weight_quantizers.44 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.21e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.down_proj_weight_quantizers.45 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.down_proj_weight_quantizers.46 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.down_proj_weight_quantizers.47 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.down_proj_weight_quantizers.48 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.down_proj_weight_quantizers.49 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.down_proj_weight_quantizers.50 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.down_proj_weight_quantizers.51 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.48e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.down_proj_weight_quantizers.52 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.down_proj_weight_quantizers.53 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.down_proj_weight_quantizers.54 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.21e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.down_proj_weight_quantizers.55 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.63e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.down_proj_weight_quantizers.56 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.15e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.down_proj_weight_quantizers.57 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.down_proj_weight_quantizers.58 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.87e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.down_proj_weight_quantizers.59 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.down_proj_weight_quantizers.60 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.down_proj_weight_quantizers.61 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.down_proj_weight_quantizers.62 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.down_proj_weight_quantizers.63 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.down_proj_weight_quantizers.64 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.down_proj_weight_quantizers.65 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.24e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.down_proj_weight_quantizers.66 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.down_proj_weight_quantizers.67 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.down_proj_weight_quantizers.68 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.down_proj_weight_quantizers.69 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.down_proj_weight_quantizers.70 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.35e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.down_proj_weight_quantizers.71 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.down_proj_weight_quantizers.72 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.35e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.down_proj_weight_quantizers.73 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.down_proj_weight_quantizers.74 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.down_proj_weight_quantizers.75 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.down_proj_weight_quantizers.76 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.down_proj_weight_quantizers.77 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.down_proj_weight_quantizers.78 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.down_proj_weight_quantizers.79 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.22e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.down_proj_weight_quantizers.80 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.down_proj_weight_quantizers.81 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.down_proj_weight_quantizers.82 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.down_proj_weight_quantizers.83 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.down_proj_weight_quantizers.84 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.down_proj_weight_quantizers.85 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.25e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.down_proj_weight_quantizers.86 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=4.30e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.down_proj_weight_quantizers.87 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.down_proj_weight_quantizers.88 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.down_proj_weight_quantizers.89 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.down_proj_weight_quantizers.90 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.down_proj_weight_quantizers.91 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.52e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.down_proj_weight_quantizers.92 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.42e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.down_proj_weight_quantizers.93 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.down_proj_weight_quantizers.94 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.down_proj_weight_quantizers.95 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.24e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.down_proj_weight_quantizers.96 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.down_proj_weight_quantizers.97 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.down_proj_weight_quantizers.98 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.25e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.down_proj_weight_quantizers.99 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.23e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.down_proj_weight_quantizers.100 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.15e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.down_proj_weight_quantizers.101 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.21e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.down_proj_weight_quantizers.102 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=6.21e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.down_proj_weight_quantizers.103 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.58e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.down_proj_weight_quantizers.104 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.60e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.down_proj_weight_quantizers.105 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.down_proj_weight_quantizers.106 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.95e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.down_proj_weight_quantizers.107 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.77e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.down_proj_weight_quantizers.108 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.down_proj_weight_quantizers.109 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.down_proj_weight_quantizers.110 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.down_proj_weight_quantizers.111 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.down_proj_weight_quantizers.112 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.down_proj_weight_quantizers.113 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.17e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.down_proj_weight_quantizers.114 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.down_proj_weight_quantizers.115 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.down_proj_weight_quantizers.116 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.21e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.down_proj_weight_quantizers.117 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.down_proj_weight_quantizers.118 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.down_proj_weight_quantizers.119 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.down_proj_weight_quantizers.120 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.28e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.down_proj_weight_quantizers.121 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.23e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.down_proj_weight_quantizers.122 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.51e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.down_proj_weight_quantizers.123 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.down_proj_weight_quantizers.124 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.63e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.down_proj_weight_quantizers.125 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.down_proj_weight_quantizers.126 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.53e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.down_proj_weight_quantizers.127 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.58e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.down_proj_weight_quantizers.128 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.down_proj_weight_quantizers.129 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.down_proj_weight_quantizers.130 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.down_proj_weight_quantizers.131 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.down_proj_weight_quantizers.132 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.18e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.down_proj_weight_quantizers.133 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.down_proj_weight_quantizers.134 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.down_proj_weight_quantizers.135 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.29e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.down_proj_weight_quantizers.136 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.down_proj_weight_quantizers.137 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.down_proj_weight_quantizers.138 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.down_proj_weight_quantizers.139 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.down_proj_weight_quantizers.140 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.93e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.down_proj_weight_quantizers.141 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.60e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.down_proj_weight_quantizers.142 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.down_proj_weight_quantizers.143 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.down_proj_weight_quantizers.144 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.48e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.down_proj_weight_quantizers.145 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.down_proj_weight_quantizers.146 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.down_proj_weight_quantizers.147 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.44e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.down_proj_weight_quantizers.148 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.down_proj_weight_quantizers.149 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.down_proj_weight_quantizers.150 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.down_proj_weight_quantizers.151 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.19e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.down_proj_weight_quantizers.152 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.down_proj_weight_quantizers.153 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.down_proj_weight_quantizers.154 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.98e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.down_proj_weight_quantizers.155 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.56e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.down_proj_weight_quantizers.156 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.39e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.down_proj_weight_quantizers.157 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.20e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.down_proj_weight_quantizers.158 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.30e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.down_proj_weight_quantizers.159 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.31e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.down_proj_weight_quantizers.160 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.28e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.down_proj_weight_quantizers.161 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.65e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.down_proj_weight_quantizers.162 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.down_proj_weight_quantizers.163 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.down_proj_weight_quantizers.164 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.down_proj_weight_quantizers.165 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.05e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.down_proj_weight_quantizers.166 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.down_proj_weight_quantizers.167 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.down_proj_weight_quantizers.168 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.down_proj_weight_quantizers.169 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.20e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.down_proj_weight_quantizers.170 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.59e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.down_proj_weight_quantizers.171 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.down_proj_weight_quantizers.172 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.66e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.down_proj_weight_quantizers.173 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.89e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.down_proj_weight_quantizers.174 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.down_proj_weight_quantizers.175 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.72e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.down_proj_weight_quantizers.176 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.down_proj_weight_quantizers.177 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.down_proj_weight_quantizers.178 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.down_proj_weight_quantizers.179 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.down_proj_weight_quantizers.180 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.64e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.down_proj_weight_quantizers.181 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.down_proj_weight_quantizers.182 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.down_proj_weight_quantizers.183 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.down_proj_weight_quantizers.184 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.44e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.down_proj_weight_quantizers.185 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.28e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.down_proj_weight_quantizers.186 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.down_proj_weight_quantizers.187 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.13e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.down_proj_weight_quantizers.188 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.down_proj_weight_quantizers.189 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.down_proj_weight_quantizers.190 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.down_proj_weight_quantizers.191 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.62e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.down_proj_weight_quantizers.192 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.down_proj_weight_quantizers.193 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.down_proj_weight_quantizers.194 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.down_proj_weight_quantizers.195 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.23e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.down_proj_weight_quantizers.196 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.down_proj_weight_quantizers.197 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.down_proj_weight_quantizers.198 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.down_proj_weight_quantizers.199 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.21e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.down_proj_weight_quantizers.200 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.40e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.down_proj_weight_quantizers.201 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.down_proj_weight_quantizers.202 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.down_proj_weight_quantizers.203 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.down_proj_weight_quantizers.204 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.down_proj_weight_quantizers.205 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.13e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.down_proj_weight_quantizers.206 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=7.62e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.down_proj_weight_quantizers.207 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.down_proj_weight_quantizers.208 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.down_proj_weight_quantizers.209 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.17e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.down_proj_weight_quantizers.210 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.69e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.down_proj_weight_quantizers.211 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.12e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.down_proj_weight_quantizers.212 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.down_proj_weight_quantizers.213 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.down_proj_weight_quantizers.214 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.down_proj_weight_quantizers.215 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.down_proj_weight_quantizers.216 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.down_proj_weight_quantizers.217 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.17e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.down_proj_weight_quantizers.218 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.72e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.down_proj_weight_quantizers.219 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.down_proj_weight_quantizers.220 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.21e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.down_proj_weight_quantizers.221 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.down_proj_weight_quantizers.222 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.down_proj_weight_quantizers.223 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.58e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.down_proj_weight_quantizers.224 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.down_proj_weight_quantizers.225 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.17e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.down_proj_weight_quantizers.226 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.57e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.down_proj_weight_quantizers.227 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.down_proj_weight_quantizers.228 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.down_proj_weight_quantizers.229 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.54e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.down_proj_weight_quantizers.230 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.down_proj_weight_quantizers.231 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.down_proj_weight_quantizers.232 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.down_proj_weight_quantizers.233 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.36e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.down_proj_weight_quantizers.234 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.17e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.down_proj_weight_quantizers.235 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.down_proj_weight_quantizers.236 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.down_proj_weight_quantizers.237 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.down_proj_weight_quantizers.238 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.down_proj_weight_quantizers.239 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.52e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.down_proj_weight_quantizers.240 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.down_proj_weight_quantizers.241 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.69e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.down_proj_weight_quantizers.242 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.42e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.down_proj_weight_quantizers.243 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.75e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.down_proj_weight_quantizers.244 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.down_proj_weight_quantizers.245 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.down_proj_weight_quantizers.246 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.down_proj_weight_quantizers.247 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.24e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.down_proj_weight_quantizers.248 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.down_proj_weight_quantizers.249 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.down_proj_weight_quantizers.250 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.down_proj_weight_quantizers.251 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.down_proj_weight_quantizers.252 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.16e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.down_proj_weight_quantizers.253 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.down_proj_weight_quantizers.254 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.93e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.down_proj_weight_quantizers.255 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.8.mlp.shared_experts.gate_proj.input_quantizer TensorQuantizer(disabled) +model.layers.8.mlp.shared_experts.gate_proj.output_quantizer TensorQuantizer(disabled) +model.layers.8.mlp.shared_experts.gate_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.8.mlp.shared_experts.up_proj.input_quantizer TensorQuantizer(disabled) +model.layers.8.mlp.shared_experts.up_proj.output_quantizer TensorQuantizer(disabled) +model.layers.8.mlp.shared_experts.up_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.8.mlp.shared_experts.down_proj.input_quantizer TensorQuantizer(disabled) +model.layers.8.mlp.shared_experts.down_proj.output_quantizer TensorQuantizer(disabled) +model.layers.8.mlp.shared_experts.down_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.9.self_attn.q_a_proj.input_quantizer TensorQuantizer(disabled) +model.layers.9.self_attn.q_a_proj.output_quantizer TensorQuantizer(disabled) +model.layers.9.self_attn.q_a_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.9.self_attn.q_b_proj.input_quantizer TensorQuantizer(disabled) +model.layers.9.self_attn.q_b_proj.output_quantizer TensorQuantizer(disabled) +model.layers.9.self_attn.q_b_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.9.self_attn.kv_a_proj_with_mqa.input_quantizer TensorQuantizer(disabled) +model.layers.9.self_attn.kv_a_proj_with_mqa.output_quantizer TensorQuantizer(disabled) +model.layers.9.self_attn.kv_a_proj_with_mqa.weight_quantizer TensorQuantizer(disabled) +model.layers.9.self_attn.kv_b_proj.input_quantizer TensorQuantizer(disabled) +model.layers.9.self_attn.kv_b_proj.output_quantizer TensorQuantizer(disabled) +model.layers.9.self_attn.kv_b_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.9.self_attn.o_proj.input_quantizer TensorQuantizer(disabled) +model.layers.9.self_attn.o_proj.output_quantizer TensorQuantizer(disabled) +model.layers.9.self_attn.o_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.9.self_attn.q_bmm_quantizer TensorQuantizer(disabled) +model.layers.9.self_attn.k_bmm_quantizer TensorQuantizer((4, 3) bit fake per-tensor amax=4.48e+02(const) calibrator=MaxCalibrator quant) +model.layers.9.self_attn.v_bmm_quantizer TensorQuantizer((4, 3) bit fake per-tensor amax=4.48e+02(const) calibrator=MaxCalibrator quant) +model.layers.9.self_attn.p_bmm_quantizer TensorQuantizer(disabled) +model.layers.9.mlp.experts.gate_up_proj_input_quantizer TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=6.09e+00 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.gate_up_proj_weight_quantizers.0 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.gate_up_proj_weight_quantizers.1 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.gate_up_proj_weight_quantizers.2 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.gate_up_proj_weight_quantizers.3 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.gate_up_proj_weight_quantizers.4 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.gate_up_proj_weight_quantizers.5 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.gate_up_proj_weight_quantizers.6 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.gate_up_proj_weight_quantizers.7 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.27e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.gate_up_proj_weight_quantizers.8 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.16e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.gate_up_proj_weight_quantizers.9 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.gate_up_proj_weight_quantizers.10 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.gate_up_proj_weight_quantizers.11 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.gate_up_proj_weight_quantizers.12 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.gate_up_proj_weight_quantizers.13 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.17e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.gate_up_proj_weight_quantizers.14 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.gate_up_proj_weight_quantizers.15 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.20e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.gate_up_proj_weight_quantizers.16 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.gate_up_proj_weight_quantizers.17 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.gate_up_proj_weight_quantizers.18 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.gate_up_proj_weight_quantizers.19 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=7.30e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.gate_up_proj_weight_quantizers.20 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.gate_up_proj_weight_quantizers.21 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.gate_up_proj_weight_quantizers.22 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.gate_up_proj_weight_quantizers.23 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.gate_up_proj_weight_quantizers.24 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.80e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.gate_up_proj_weight_quantizers.25 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.gate_up_proj_weight_quantizers.26 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.gate_up_proj_weight_quantizers.27 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.gate_up_proj_weight_quantizers.28 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.gate_up_proj_weight_quantizers.29 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.gate_up_proj_weight_quantizers.30 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.gate_up_proj_weight_quantizers.31 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.39e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.gate_up_proj_weight_quantizers.32 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.gate_up_proj_weight_quantizers.33 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.46e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.gate_up_proj_weight_quantizers.34 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.gate_up_proj_weight_quantizers.35 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.gate_up_proj_weight_quantizers.36 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.gate_up_proj_weight_quantizers.37 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.gate_up_proj_weight_quantizers.38 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.gate_up_proj_weight_quantizers.39 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.gate_up_proj_weight_quantizers.40 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.gate_up_proj_weight_quantizers.41 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.gate_up_proj_weight_quantizers.42 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.gate_up_proj_weight_quantizers.43 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.gate_up_proj_weight_quantizers.44 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.gate_up_proj_weight_quantizers.45 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.13e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.gate_up_proj_weight_quantizers.46 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.gate_up_proj_weight_quantizers.47 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.gate_up_proj_weight_quantizers.48 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.17e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.gate_up_proj_weight_quantizers.49 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.82e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.gate_up_proj_weight_quantizers.50 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.gate_up_proj_weight_quantizers.51 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.14e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.gate_up_proj_weight_quantizers.52 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.42e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.gate_up_proj_weight_quantizers.53 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.gate_up_proj_weight_quantizers.54 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.gate_up_proj_weight_quantizers.55 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.gate_up_proj_weight_quantizers.56 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.gate_up_proj_weight_quantizers.57 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.gate_up_proj_weight_quantizers.58 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.15e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.gate_up_proj_weight_quantizers.59 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.gate_up_proj_weight_quantizers.60 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.15e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.gate_up_proj_weight_quantizers.61 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.gate_up_proj_weight_quantizers.62 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.gate_up_proj_weight_quantizers.63 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.gate_up_proj_weight_quantizers.64 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.gate_up_proj_weight_quantizers.65 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.gate_up_proj_weight_quantizers.66 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.46e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.gate_up_proj_weight_quantizers.67 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.gate_up_proj_weight_quantizers.68 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.gate_up_proj_weight_quantizers.69 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.gate_up_proj_weight_quantizers.70 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.gate_up_proj_weight_quantizers.71 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.gate_up_proj_weight_quantizers.72 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.gate_up_proj_weight_quantizers.73 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.97e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.gate_up_proj_weight_quantizers.74 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.gate_up_proj_weight_quantizers.75 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.gate_up_proj_weight_quantizers.76 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.gate_up_proj_weight_quantizers.77 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.gate_up_proj_weight_quantizers.78 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.16e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.gate_up_proj_weight_quantizers.79 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.gate_up_proj_weight_quantizers.80 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.gate_up_proj_weight_quantizers.81 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.gate_up_proj_weight_quantizers.82 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.gate_up_proj_weight_quantizers.83 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.gate_up_proj_weight_quantizers.84 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.00e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.gate_up_proj_weight_quantizers.85 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.gate_up_proj_weight_quantizers.86 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.gate_up_proj_weight_quantizers.87 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.gate_up_proj_weight_quantizers.88 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.gate_up_proj_weight_quantizers.89 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.gate_up_proj_weight_quantizers.90 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.gate_up_proj_weight_quantizers.91 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.gate_up_proj_weight_quantizers.92 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.50e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.gate_up_proj_weight_quantizers.93 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.gate_up_proj_weight_quantizers.94 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.gate_up_proj_weight_quantizers.95 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.gate_up_proj_weight_quantizers.96 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.gate_up_proj_weight_quantizers.97 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.gate_up_proj_weight_quantizers.98 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.gate_up_proj_weight_quantizers.99 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.gate_up_proj_weight_quantizers.100 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.09e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.gate_up_proj_weight_quantizers.101 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.15e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.gate_up_proj_weight_quantizers.102 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.gate_up_proj_weight_quantizers.103 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.gate_up_proj_weight_quantizers.104 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.gate_up_proj_weight_quantizers.105 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.53e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.gate_up_proj_weight_quantizers.106 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.gate_up_proj_weight_quantizers.107 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.gate_up_proj_weight_quantizers.108 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.gate_up_proj_weight_quantizers.109 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.gate_up_proj_weight_quantizers.110 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.gate_up_proj_weight_quantizers.111 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.64e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.gate_up_proj_weight_quantizers.112 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.gate_up_proj_weight_quantizers.113 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.gate_up_proj_weight_quantizers.114 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.gate_up_proj_weight_quantizers.115 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.gate_up_proj_weight_quantizers.116 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.gate_up_proj_weight_quantizers.117 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.gate_up_proj_weight_quantizers.118 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.gate_up_proj_weight_quantizers.119 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.50e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.gate_up_proj_weight_quantizers.120 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.gate_up_proj_weight_quantizers.121 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.gate_up_proj_weight_quantizers.122 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.21e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.gate_up_proj_weight_quantizers.123 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.gate_up_proj_weight_quantizers.124 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.gate_up_proj_weight_quantizers.125 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.gate_up_proj_weight_quantizers.126 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.gate_up_proj_weight_quantizers.127 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.gate_up_proj_weight_quantizers.128 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.gate_up_proj_weight_quantizers.129 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.gate_up_proj_weight_quantizers.130 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.gate_up_proj_weight_quantizers.131 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=6.25e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.gate_up_proj_weight_quantizers.132 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.gate_up_proj_weight_quantizers.133 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.gate_up_proj_weight_quantizers.134 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.gate_up_proj_weight_quantizers.135 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.gate_up_proj_weight_quantizers.136 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.gate_up_proj_weight_quantizers.137 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.gate_up_proj_weight_quantizers.138 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.gate_up_proj_weight_quantizers.139 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.gate_up_proj_weight_quantizers.140 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.gate_up_proj_weight_quantizers.141 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.gate_up_proj_weight_quantizers.142 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.gate_up_proj_weight_quantizers.143 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.gate_up_proj_weight_quantizers.144 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.gate_up_proj_weight_quantizers.145 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.50e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.gate_up_proj_weight_quantizers.146 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.gate_up_proj_weight_quantizers.147 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.gate_up_proj_weight_quantizers.148 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.gate_up_proj_weight_quantizers.149 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.gate_up_proj_weight_quantizers.150 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.gate_up_proj_weight_quantizers.151 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.10e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.gate_up_proj_weight_quantizers.152 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.gate_up_proj_weight_quantizers.153 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.gate_up_proj_weight_quantizers.154 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.gate_up_proj_weight_quantizers.155 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.gate_up_proj_weight_quantizers.156 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.gate_up_proj_weight_quantizers.157 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.gate_up_proj_weight_quantizers.158 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.gate_up_proj_weight_quantizers.159 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.gate_up_proj_weight_quantizers.160 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.14e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.gate_up_proj_weight_quantizers.161 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.gate_up_proj_weight_quantizers.162 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.gate_up_proj_weight_quantizers.163 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.gate_up_proj_weight_quantizers.164 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.gate_up_proj_weight_quantizers.165 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.gate_up_proj_weight_quantizers.166 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.gate_up_proj_weight_quantizers.167 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.gate_up_proj_weight_quantizers.168 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.gate_up_proj_weight_quantizers.169 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.gate_up_proj_weight_quantizers.170 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.36e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.gate_up_proj_weight_quantizers.171 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.13e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.gate_up_proj_weight_quantizers.172 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.gate_up_proj_weight_quantizers.173 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.gate_up_proj_weight_quantizers.174 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.gate_up_proj_weight_quantizers.175 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.gate_up_proj_weight_quantizers.176 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.gate_up_proj_weight_quantizers.177 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.56e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.gate_up_proj_weight_quantizers.178 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.gate_up_proj_weight_quantizers.179 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.gate_up_proj_weight_quantizers.180 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.gate_up_proj_weight_quantizers.181 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.gate_up_proj_weight_quantizers.182 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.gate_up_proj_weight_quantizers.183 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.gate_up_proj_weight_quantizers.184 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.gate_up_proj_weight_quantizers.185 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.54e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.gate_up_proj_weight_quantizers.186 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.gate_up_proj_weight_quantizers.187 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.gate_up_proj_weight_quantizers.188 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.gate_up_proj_weight_quantizers.189 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.gate_up_proj_weight_quantizers.190 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.gate_up_proj_weight_quantizers.191 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.19e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.gate_up_proj_weight_quantizers.192 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.gate_up_proj_weight_quantizers.193 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.14e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.gate_up_proj_weight_quantizers.194 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.13e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.gate_up_proj_weight_quantizers.195 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.gate_up_proj_weight_quantizers.196 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.gate_up_proj_weight_quantizers.197 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.gate_up_proj_weight_quantizers.198 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.gate_up_proj_weight_quantizers.199 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.gate_up_proj_weight_quantizers.200 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.gate_up_proj_weight_quantizers.201 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.gate_up_proj_weight_quantizers.202 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.14e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.gate_up_proj_weight_quantizers.203 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.gate_up_proj_weight_quantizers.204 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.36e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.gate_up_proj_weight_quantizers.205 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.gate_up_proj_weight_quantizers.206 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.gate_up_proj_weight_quantizers.207 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.13e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.gate_up_proj_weight_quantizers.208 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.gate_up_proj_weight_quantizers.209 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.gate_up_proj_weight_quantizers.210 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.90e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.gate_up_proj_weight_quantizers.211 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.gate_up_proj_weight_quantizers.212 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.gate_up_proj_weight_quantizers.213 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.gate_up_proj_weight_quantizers.214 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.gate_up_proj_weight_quantizers.215 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.gate_up_proj_weight_quantizers.216 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.gate_up_proj_weight_quantizers.217 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.gate_up_proj_weight_quantizers.218 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.gate_up_proj_weight_quantizers.219 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.gate_up_proj_weight_quantizers.220 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.gate_up_proj_weight_quantizers.221 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.gate_up_proj_weight_quantizers.222 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.gate_up_proj_weight_quantizers.223 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.gate_up_proj_weight_quantizers.224 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.gate_up_proj_weight_quantizers.225 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.62e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.gate_up_proj_weight_quantizers.226 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.gate_up_proj_weight_quantizers.227 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.gate_up_proj_weight_quantizers.228 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.41e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.gate_up_proj_weight_quantizers.229 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.gate_up_proj_weight_quantizers.230 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.gate_up_proj_weight_quantizers.231 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.gate_up_proj_weight_quantizers.232 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.gate_up_proj_weight_quantizers.233 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.gate_up_proj_weight_quantizers.234 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.gate_up_proj_weight_quantizers.235 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.gate_up_proj_weight_quantizers.236 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.81e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.gate_up_proj_weight_quantizers.237 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.gate_up_proj_weight_quantizers.238 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.gate_up_proj_weight_quantizers.239 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.gate_up_proj_weight_quantizers.240 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.gate_up_proj_weight_quantizers.241 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.gate_up_proj_weight_quantizers.242 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.16e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.gate_up_proj_weight_quantizers.243 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.36e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.gate_up_proj_weight_quantizers.244 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.gate_up_proj_weight_quantizers.245 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.gate_up_proj_weight_quantizers.246 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.gate_up_proj_weight_quantizers.247 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.gate_up_proj_weight_quantizers.248 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.gate_up_proj_weight_quantizers.249 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.gate_up_proj_weight_quantizers.250 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.gate_up_proj_weight_quantizers.251 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.gate_up_proj_weight_quantizers.252 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.gate_up_proj_weight_quantizers.253 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.gate_up_proj_weight_quantizers.254 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.gate_up_proj_weight_quantizers.255 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.down_proj_input_quantizer TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.00e+00 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.down_proj_weight_quantizers.0 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.down_proj_weight_quantizers.1 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.down_proj_weight_quantizers.2 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.89e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.down_proj_weight_quantizers.3 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.down_proj_weight_quantizers.4 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.49e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.down_proj_weight_quantizers.5 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.down_proj_weight_quantizers.6 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.19e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.down_proj_weight_quantizers.7 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.15e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.down_proj_weight_quantizers.8 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.20e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.down_proj_weight_quantizers.9 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.13e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.down_proj_weight_quantizers.10 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.down_proj_weight_quantizers.11 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.down_proj_weight_quantizers.12 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.38e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.down_proj_weight_quantizers.13 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.25e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.down_proj_weight_quantizers.14 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.down_proj_weight_quantizers.15 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.75e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.down_proj_weight_quantizers.16 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.down_proj_weight_quantizers.17 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.down_proj_weight_quantizers.18 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.down_proj_weight_quantizers.19 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.14e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.down_proj_weight_quantizers.20 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.down_proj_weight_quantizers.21 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.18e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.down_proj_weight_quantizers.22 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.15e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.down_proj_weight_quantizers.23 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.21e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.down_proj_weight_quantizers.24 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.26e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.down_proj_weight_quantizers.25 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.down_proj_weight_quantizers.26 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.down_proj_weight_quantizers.27 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.14e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.down_proj_weight_quantizers.28 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.down_proj_weight_quantizers.29 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.down_proj_weight_quantizers.30 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.29e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.down_proj_weight_quantizers.31 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.down_proj_weight_quantizers.32 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.down_proj_weight_quantizers.33 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.down_proj_weight_quantizers.34 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.down_proj_weight_quantizers.35 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.down_proj_weight_quantizers.36 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.68e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.down_proj_weight_quantizers.37 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.down_proj_weight_quantizers.38 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.49e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.down_proj_weight_quantizers.39 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.down_proj_weight_quantizers.40 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.97e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.down_proj_weight_quantizers.41 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.down_proj_weight_quantizers.42 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.down_proj_weight_quantizers.43 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.down_proj_weight_quantizers.44 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.down_proj_weight_quantizers.45 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.down_proj_weight_quantizers.46 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.down_proj_weight_quantizers.47 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.down_proj_weight_quantizers.48 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.down_proj_weight_quantizers.49 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.41e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.down_proj_weight_quantizers.50 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.down_proj_weight_quantizers.51 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.41e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.down_proj_weight_quantizers.52 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.46e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.down_proj_weight_quantizers.53 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.33e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.down_proj_weight_quantizers.54 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.56e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.down_proj_weight_quantizers.55 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.down_proj_weight_quantizers.56 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.down_proj_weight_quantizers.57 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.down_proj_weight_quantizers.58 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.20e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.down_proj_weight_quantizers.59 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.down_proj_weight_quantizers.60 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.down_proj_weight_quantizers.61 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.down_proj_weight_quantizers.62 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.18e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.down_proj_weight_quantizers.63 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.down_proj_weight_quantizers.64 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.down_proj_weight_quantizers.65 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.down_proj_weight_quantizers.66 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.down_proj_weight_quantizers.67 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.down_proj_weight_quantizers.68 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.down_proj_weight_quantizers.69 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.13e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.down_proj_weight_quantizers.70 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.down_proj_weight_quantizers.71 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.down_proj_weight_quantizers.72 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.14e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.down_proj_weight_quantizers.73 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.down_proj_weight_quantizers.74 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.17e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.down_proj_weight_quantizers.75 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.39e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.down_proj_weight_quantizers.76 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.down_proj_weight_quantizers.77 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.down_proj_weight_quantizers.78 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.50e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.down_proj_weight_quantizers.79 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.down_proj_weight_quantizers.80 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.down_proj_weight_quantizers.81 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.27e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.down_proj_weight_quantizers.82 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.down_proj_weight_quantizers.83 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.down_proj_weight_quantizers.84 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.27e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.down_proj_weight_quantizers.85 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.down_proj_weight_quantizers.86 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.down_proj_weight_quantizers.87 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.down_proj_weight_quantizers.88 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.down_proj_weight_quantizers.89 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.down_proj_weight_quantizers.90 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.26e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.down_proj_weight_quantizers.91 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.20e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.down_proj_weight_quantizers.92 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.21e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.down_proj_weight_quantizers.93 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.down_proj_weight_quantizers.94 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.36e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.down_proj_weight_quantizers.95 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.down_proj_weight_quantizers.96 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.down_proj_weight_quantizers.97 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.down_proj_weight_quantizers.98 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.down_proj_weight_quantizers.99 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.down_proj_weight_quantizers.100 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.down_proj_weight_quantizers.101 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.45e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.down_proj_weight_quantizers.102 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.33e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.down_proj_weight_quantizers.103 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.down_proj_weight_quantizers.104 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.down_proj_weight_quantizers.105 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.21e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.down_proj_weight_quantizers.106 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.down_proj_weight_quantizers.107 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.down_proj_weight_quantizers.108 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.29e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.down_proj_weight_quantizers.109 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.down_proj_weight_quantizers.110 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.24e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.down_proj_weight_quantizers.111 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.down_proj_weight_quantizers.112 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.down_proj_weight_quantizers.113 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.down_proj_weight_quantizers.114 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.down_proj_weight_quantizers.115 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.down_proj_weight_quantizers.116 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.69e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.down_proj_weight_quantizers.117 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.down_proj_weight_quantizers.118 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.down_proj_weight_quantizers.119 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.63e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.down_proj_weight_quantizers.120 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.down_proj_weight_quantizers.121 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.28e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.down_proj_weight_quantizers.122 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.down_proj_weight_quantizers.123 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.down_proj_weight_quantizers.124 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.down_proj_weight_quantizers.125 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.down_proj_weight_quantizers.126 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.down_proj_weight_quantizers.127 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.down_proj_weight_quantizers.128 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.28e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.down_proj_weight_quantizers.129 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.down_proj_weight_quantizers.130 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.down_proj_weight_quantizers.131 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.91e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.down_proj_weight_quantizers.132 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.50e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.down_proj_weight_quantizers.133 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.down_proj_weight_quantizers.134 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.down_proj_weight_quantizers.135 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.down_proj_weight_quantizers.136 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.down_proj_weight_quantizers.137 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.down_proj_weight_quantizers.138 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.32e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.down_proj_weight_quantizers.139 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.down_proj_weight_quantizers.140 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.81e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.down_proj_weight_quantizers.141 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.21e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.down_proj_weight_quantizers.142 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.down_proj_weight_quantizers.143 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.down_proj_weight_quantizers.144 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.46e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.down_proj_weight_quantizers.145 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.down_proj_weight_quantizers.146 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.28e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.down_proj_weight_quantizers.147 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.down_proj_weight_quantizers.148 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.down_proj_weight_quantizers.149 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.40e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.down_proj_weight_quantizers.150 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.down_proj_weight_quantizers.151 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.94e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.down_proj_weight_quantizers.152 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.down_proj_weight_quantizers.153 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.68e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.down_proj_weight_quantizers.154 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.44e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.down_proj_weight_quantizers.155 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.down_proj_weight_quantizers.156 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.down_proj_weight_quantizers.157 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.down_proj_weight_quantizers.158 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.down_proj_weight_quantizers.159 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.24e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.down_proj_weight_quantizers.160 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.down_proj_weight_quantizers.161 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.down_proj_weight_quantizers.162 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.down_proj_weight_quantizers.163 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.down_proj_weight_quantizers.164 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.14e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.down_proj_weight_quantizers.165 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.27e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.down_proj_weight_quantizers.166 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.19e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.down_proj_weight_quantizers.167 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.19e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.down_proj_weight_quantizers.168 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.down_proj_weight_quantizers.169 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.down_proj_weight_quantizers.170 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.49e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.down_proj_weight_quantizers.171 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.23e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.down_proj_weight_quantizers.172 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.down_proj_weight_quantizers.173 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.73e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.down_proj_weight_quantizers.174 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.down_proj_weight_quantizers.175 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.down_proj_weight_quantizers.176 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.down_proj_weight_quantizers.177 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.45e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.down_proj_weight_quantizers.178 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.down_proj_weight_quantizers.179 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.down_proj_weight_quantizers.180 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.down_proj_weight_quantizers.181 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.down_proj_weight_quantizers.182 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.down_proj_weight_quantizers.183 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.down_proj_weight_quantizers.184 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.down_proj_weight_quantizers.185 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.down_proj_weight_quantizers.186 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.down_proj_weight_quantizers.187 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.down_proj_weight_quantizers.188 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.36e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.down_proj_weight_quantizers.189 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.down_proj_weight_quantizers.190 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.down_proj_weight_quantizers.191 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.55e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.down_proj_weight_quantizers.192 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.down_proj_weight_quantizers.193 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.down_proj_weight_quantizers.194 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.38e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.down_proj_weight_quantizers.195 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.down_proj_weight_quantizers.196 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.13e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.down_proj_weight_quantizers.197 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.49e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.down_proj_weight_quantizers.198 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.down_proj_weight_quantizers.199 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.64e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.down_proj_weight_quantizers.200 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.down_proj_weight_quantizers.201 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.down_proj_weight_quantizers.202 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.24e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.down_proj_weight_quantizers.203 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.14e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.down_proj_weight_quantizers.204 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.down_proj_weight_quantizers.205 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.down_proj_weight_quantizers.206 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.19e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.down_proj_weight_quantizers.207 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.13e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.down_proj_weight_quantizers.208 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.down_proj_weight_quantizers.209 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.25e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.down_proj_weight_quantizers.210 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.down_proj_weight_quantizers.211 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.down_proj_weight_quantizers.212 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.down_proj_weight_quantizers.213 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.down_proj_weight_quantizers.214 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.down_proj_weight_quantizers.215 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.down_proj_weight_quantizers.216 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.down_proj_weight_quantizers.217 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.down_proj_weight_quantizers.218 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.33e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.down_proj_weight_quantizers.219 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.down_proj_weight_quantizers.220 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.78e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.down_proj_weight_quantizers.221 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.down_proj_weight_quantizers.222 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.41e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.down_proj_weight_quantizers.223 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.26e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.down_proj_weight_quantizers.224 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.14e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.down_proj_weight_quantizers.225 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.30e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.down_proj_weight_quantizers.226 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.down_proj_weight_quantizers.227 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.39e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.down_proj_weight_quantizers.228 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.25e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.down_proj_weight_quantizers.229 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.down_proj_weight_quantizers.230 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.27e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.down_proj_weight_quantizers.231 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.down_proj_weight_quantizers.232 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.down_proj_weight_quantizers.233 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.down_proj_weight_quantizers.234 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.down_proj_weight_quantizers.235 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.down_proj_weight_quantizers.236 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.25e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.down_proj_weight_quantizers.237 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.down_proj_weight_quantizers.238 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.down_proj_weight_quantizers.239 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.27e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.down_proj_weight_quantizers.240 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.down_proj_weight_quantizers.241 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.down_proj_weight_quantizers.242 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.46e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.down_proj_weight_quantizers.243 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.83e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.down_proj_weight_quantizers.244 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.down_proj_weight_quantizers.245 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.down_proj_weight_quantizers.246 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.down_proj_weight_quantizers.247 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.down_proj_weight_quantizers.248 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.19e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.down_proj_weight_quantizers.249 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.down_proj_weight_quantizers.250 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.down_proj_weight_quantizers.251 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.down_proj_weight_quantizers.252 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.down_proj_weight_quantizers.253 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.down_proj_weight_quantizers.254 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.16e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.down_proj_weight_quantizers.255 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.9.mlp.shared_experts.gate_proj.input_quantizer TensorQuantizer(disabled) +model.layers.9.mlp.shared_experts.gate_proj.output_quantizer TensorQuantizer(disabled) +model.layers.9.mlp.shared_experts.gate_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.9.mlp.shared_experts.up_proj.input_quantizer TensorQuantizer(disabled) +model.layers.9.mlp.shared_experts.up_proj.output_quantizer TensorQuantizer(disabled) +model.layers.9.mlp.shared_experts.up_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.9.mlp.shared_experts.down_proj.input_quantizer TensorQuantizer(disabled) +model.layers.9.mlp.shared_experts.down_proj.output_quantizer TensorQuantizer(disabled) +model.layers.9.mlp.shared_experts.down_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.10.self_attn.q_a_proj.input_quantizer TensorQuantizer(disabled) +model.layers.10.self_attn.q_a_proj.output_quantizer TensorQuantizer(disabled) +model.layers.10.self_attn.q_a_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.10.self_attn.q_b_proj.input_quantizer TensorQuantizer(disabled) +model.layers.10.self_attn.q_b_proj.output_quantizer TensorQuantizer(disabled) +model.layers.10.self_attn.q_b_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.10.self_attn.kv_a_proj_with_mqa.input_quantizer TensorQuantizer(disabled) +model.layers.10.self_attn.kv_a_proj_with_mqa.output_quantizer TensorQuantizer(disabled) +model.layers.10.self_attn.kv_a_proj_with_mqa.weight_quantizer TensorQuantizer(disabled) +model.layers.10.self_attn.kv_b_proj.input_quantizer TensorQuantizer(disabled) +model.layers.10.self_attn.kv_b_proj.output_quantizer TensorQuantizer(disabled) +model.layers.10.self_attn.kv_b_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.10.self_attn.o_proj.input_quantizer TensorQuantizer(disabled) +model.layers.10.self_attn.o_proj.output_quantizer TensorQuantizer(disabled) +model.layers.10.self_attn.o_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.10.self_attn.indexer.wq_b.input_quantizer TensorQuantizer(disabled) +model.layers.10.self_attn.indexer.wq_b.output_quantizer TensorQuantizer(disabled) +model.layers.10.self_attn.indexer.wq_b.weight_quantizer TensorQuantizer(disabled) +model.layers.10.self_attn.indexer.wk.input_quantizer TensorQuantizer(disabled) +model.layers.10.self_attn.indexer.wk.output_quantizer TensorQuantizer(disabled) +model.layers.10.self_attn.indexer.wk.weight_quantizer TensorQuantizer(disabled) +model.layers.10.self_attn.indexer.k_norm.input_quantizer TensorQuantizer(disabled) +model.layers.10.self_attn.indexer.k_norm.output_quantizer TensorQuantizer(disabled) +model.layers.10.self_attn.indexer.weights_proj.input_quantizer TensorQuantizer(disabled) +model.layers.10.self_attn.indexer.weights_proj.output_quantizer TensorQuantizer(disabled) +model.layers.10.self_attn.indexer.weights_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.10.self_attn.q_bmm_quantizer TensorQuantizer(disabled) +model.layers.10.self_attn.k_bmm_quantizer TensorQuantizer((4, 3) bit fake per-tensor amax=4.48e+02(const) calibrator=MaxCalibrator quant) +model.layers.10.self_attn.v_bmm_quantizer TensorQuantizer((4, 3) bit fake per-tensor amax=4.48e+02(const) calibrator=MaxCalibrator quant) +model.layers.10.self_attn.p_bmm_quantizer TensorQuantizer(disabled) +model.layers.10.mlp.experts.gate_up_proj_input_quantizer TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=4.53e-01 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.gate_up_proj_weight_quantizers.0 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.gate_up_proj_weight_quantizers.1 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.gate_up_proj_weight_quantizers.2 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.gate_up_proj_weight_quantizers.3 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.gate_up_proj_weight_quantizers.4 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.gate_up_proj_weight_quantizers.5 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.gate_up_proj_weight_quantizers.6 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.gate_up_proj_weight_quantizers.7 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.gate_up_proj_weight_quantizers.8 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.gate_up_proj_weight_quantizers.9 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.gate_up_proj_weight_quantizers.10 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.gate_up_proj_weight_quantizers.11 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.gate_up_proj_weight_quantizers.12 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.gate_up_proj_weight_quantizers.13 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.gate_up_proj_weight_quantizers.14 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.gate_up_proj_weight_quantizers.15 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.gate_up_proj_weight_quantizers.16 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.gate_up_proj_weight_quantizers.17 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.gate_up_proj_weight_quantizers.18 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.gate_up_proj_weight_quantizers.19 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.gate_up_proj_weight_quantizers.20 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.gate_up_proj_weight_quantizers.21 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.gate_up_proj_weight_quantizers.22 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.gate_up_proj_weight_quantizers.23 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.gate_up_proj_weight_quantizers.24 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.gate_up_proj_weight_quantizers.25 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.gate_up_proj_weight_quantizers.26 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.gate_up_proj_weight_quantizers.27 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.gate_up_proj_weight_quantizers.28 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.gate_up_proj_weight_quantizers.29 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.gate_up_proj_weight_quantizers.30 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.gate_up_proj_weight_quantizers.31 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.gate_up_proj_weight_quantizers.32 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.gate_up_proj_weight_quantizers.33 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.gate_up_proj_weight_quantizers.34 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.gate_up_proj_weight_quantizers.35 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.gate_up_proj_weight_quantizers.36 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.gate_up_proj_weight_quantizers.37 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.gate_up_proj_weight_quantizers.38 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.gate_up_proj_weight_quantizers.39 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.gate_up_proj_weight_quantizers.40 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.gate_up_proj_weight_quantizers.41 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.gate_up_proj_weight_quantizers.42 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.gate_up_proj_weight_quantizers.43 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.gate_up_proj_weight_quantizers.44 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.gate_up_proj_weight_quantizers.45 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.gate_up_proj_weight_quantizers.46 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.gate_up_proj_weight_quantizers.47 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.gate_up_proj_weight_quantizers.48 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.gate_up_proj_weight_quantizers.49 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.14e-01 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.gate_up_proj_weight_quantizers.50 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.gate_up_proj_weight_quantizers.51 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.gate_up_proj_weight_quantizers.52 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.gate_up_proj_weight_quantizers.53 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.gate_up_proj_weight_quantizers.54 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.gate_up_proj_weight_quantizers.55 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.gate_up_proj_weight_quantizers.56 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.gate_up_proj_weight_quantizers.57 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.gate_up_proj_weight_quantizers.58 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.gate_up_proj_weight_quantizers.59 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.gate_up_proj_weight_quantizers.60 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.gate_up_proj_weight_quantizers.61 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.gate_up_proj_weight_quantizers.62 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.gate_up_proj_weight_quantizers.63 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.gate_up_proj_weight_quantizers.64 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.25e-01 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.gate_up_proj_weight_quantizers.65 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.gate_up_proj_weight_quantizers.66 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.gate_up_proj_weight_quantizers.67 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.gate_up_proj_weight_quantizers.68 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.gate_up_proj_weight_quantizers.69 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.gate_up_proj_weight_quantizers.70 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.gate_up_proj_weight_quantizers.71 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.46e-01 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.gate_up_proj_weight_quantizers.72 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.gate_up_proj_weight_quantizers.73 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.gate_up_proj_weight_quantizers.74 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.gate_up_proj_weight_quantizers.75 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.gate_up_proj_weight_quantizers.76 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.gate_up_proj_weight_quantizers.77 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.gate_up_proj_weight_quantizers.78 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.16e-01 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.gate_up_proj_weight_quantizers.79 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.gate_up_proj_weight_quantizers.80 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.gate_up_proj_weight_quantizers.81 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.gate_up_proj_weight_quantizers.82 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.gate_up_proj_weight_quantizers.83 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.gate_up_proj_weight_quantizers.84 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.gate_up_proj_weight_quantizers.85 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.gate_up_proj_weight_quantizers.86 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.gate_up_proj_weight_quantizers.87 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.47e-01 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.gate_up_proj_weight_quantizers.88 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.gate_up_proj_weight_quantizers.89 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.gate_up_proj_weight_quantizers.90 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.gate_up_proj_weight_quantizers.91 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.gate_up_proj_weight_quantizers.92 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.gate_up_proj_weight_quantizers.93 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.gate_up_proj_weight_quantizers.94 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.gate_up_proj_weight_quantizers.95 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.35e-01 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.gate_up_proj_weight_quantizers.96 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.gate_up_proj_weight_quantizers.97 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.gate_up_proj_weight_quantizers.98 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.gate_up_proj_weight_quantizers.99 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.37e-01 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.gate_up_proj_weight_quantizers.100 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.gate_up_proj_weight_quantizers.101 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.gate_up_proj_weight_quantizers.102 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.gate_up_proj_weight_quantizers.103 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.gate_up_proj_weight_quantizers.104 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.gate_up_proj_weight_quantizers.105 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.gate_up_proj_weight_quantizers.106 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.gate_up_proj_weight_quantizers.107 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.gate_up_proj_weight_quantizers.108 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.gate_up_proj_weight_quantizers.109 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.gate_up_proj_weight_quantizers.110 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.gate_up_proj_weight_quantizers.111 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.gate_up_proj_weight_quantizers.112 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.gate_up_proj_weight_quantizers.113 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.gate_up_proj_weight_quantizers.114 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.16e-01 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.gate_up_proj_weight_quantizers.115 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.14e-01 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.gate_up_proj_weight_quantizers.116 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.gate_up_proj_weight_quantizers.117 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.gate_up_proj_weight_quantizers.118 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.gate_up_proj_weight_quantizers.119 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.gate_up_proj_weight_quantizers.120 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.gate_up_proj_weight_quantizers.121 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.gate_up_proj_weight_quantizers.122 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.gate_up_proj_weight_quantizers.123 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.gate_up_proj_weight_quantizers.124 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.gate_up_proj_weight_quantizers.125 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.gate_up_proj_weight_quantizers.126 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.15e-01 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.gate_up_proj_weight_quantizers.127 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.gate_up_proj_weight_quantizers.128 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.gate_up_proj_weight_quantizers.129 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.gate_up_proj_weight_quantizers.130 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.gate_up_proj_weight_quantizers.131 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.gate_up_proj_weight_quantizers.132 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.gate_up_proj_weight_quantizers.133 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.gate_up_proj_weight_quantizers.134 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.gate_up_proj_weight_quantizers.135 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.gate_up_proj_weight_quantizers.136 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.gate_up_proj_weight_quantizers.137 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.gate_up_proj_weight_quantizers.138 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.gate_up_proj_weight_quantizers.139 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.gate_up_proj_weight_quantizers.140 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.13e-01 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.gate_up_proj_weight_quantizers.141 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.gate_up_proj_weight_quantizers.142 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.gate_up_proj_weight_quantizers.143 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.gate_up_proj_weight_quantizers.144 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.gate_up_proj_weight_quantizers.145 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.19e-01 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.gate_up_proj_weight_quantizers.146 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.gate_up_proj_weight_quantizers.147 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.gate_up_proj_weight_quantizers.148 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.gate_up_proj_weight_quantizers.149 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.27e-01 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.gate_up_proj_weight_quantizers.150 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.gate_up_proj_weight_quantizers.151 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.gate_up_proj_weight_quantizers.152 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.gate_up_proj_weight_quantizers.153 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.gate_up_proj_weight_quantizers.154 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.gate_up_proj_weight_quantizers.155 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.gate_up_proj_weight_quantizers.156 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.gate_up_proj_weight_quantizers.157 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.gate_up_proj_weight_quantizers.158 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.gate_up_proj_weight_quantizers.159 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.gate_up_proj_weight_quantizers.160 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.gate_up_proj_weight_quantizers.161 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.gate_up_proj_weight_quantizers.162 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.17e-01 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.gate_up_proj_weight_quantizers.163 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.gate_up_proj_weight_quantizers.164 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.gate_up_proj_weight_quantizers.165 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.gate_up_proj_weight_quantizers.166 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.gate_up_proj_weight_quantizers.167 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.15e-01 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.gate_up_proj_weight_quantizers.168 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.gate_up_proj_weight_quantizers.169 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.gate_up_proj_weight_quantizers.170 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.gate_up_proj_weight_quantizers.171 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.41e-01 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.gate_up_proj_weight_quantizers.172 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.gate_up_proj_weight_quantizers.173 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.19e+00 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.gate_up_proj_weight_quantizers.174 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.gate_up_proj_weight_quantizers.175 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.gate_up_proj_weight_quantizers.176 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.gate_up_proj_weight_quantizers.177 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.gate_up_proj_weight_quantizers.178 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.gate_up_proj_weight_quantizers.179 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.gate_up_proj_weight_quantizers.180 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.gate_up_proj_weight_quantizers.181 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.gate_up_proj_weight_quantizers.182 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.gate_up_proj_weight_quantizers.183 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.gate_up_proj_weight_quantizers.184 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.17e-01 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.gate_up_proj_weight_quantizers.185 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.gate_up_proj_weight_quantizers.186 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.gate_up_proj_weight_quantizers.187 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.16e-01 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.gate_up_proj_weight_quantizers.188 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.44e-01 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.gate_up_proj_weight_quantizers.189 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.gate_up_proj_weight_quantizers.190 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.gate_up_proj_weight_quantizers.191 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.gate_up_proj_weight_quantizers.192 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.gate_up_proj_weight_quantizers.193 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.gate_up_proj_weight_quantizers.194 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.gate_up_proj_weight_quantizers.195 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.gate_up_proj_weight_quantizers.196 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.gate_up_proj_weight_quantizers.197 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.gate_up_proj_weight_quantizers.198 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.gate_up_proj_weight_quantizers.199 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.gate_up_proj_weight_quantizers.200 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.gate_up_proj_weight_quantizers.201 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.gate_up_proj_weight_quantizers.202 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.gate_up_proj_weight_quantizers.203 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.gate_up_proj_weight_quantizers.204 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.gate_up_proj_weight_quantizers.205 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.gate_up_proj_weight_quantizers.206 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.gate_up_proj_weight_quantizers.207 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.gate_up_proj_weight_quantizers.208 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.gate_up_proj_weight_quantizers.209 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.gate_up_proj_weight_quantizers.210 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.gate_up_proj_weight_quantizers.211 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.gate_up_proj_weight_quantizers.212 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.gate_up_proj_weight_quantizers.213 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.gate_up_proj_weight_quantizers.214 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.gate_up_proj_weight_quantizers.215 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.gate_up_proj_weight_quantizers.216 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.gate_up_proj_weight_quantizers.217 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.gate_up_proj_weight_quantizers.218 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.gate_up_proj_weight_quantizers.219 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.gate_up_proj_weight_quantizers.220 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.gate_up_proj_weight_quantizers.221 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.gate_up_proj_weight_quantizers.222 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.36e-01 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.gate_up_proj_weight_quantizers.223 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.gate_up_proj_weight_quantizers.224 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.gate_up_proj_weight_quantizers.225 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.gate_up_proj_weight_quantizers.226 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.gate_up_proj_weight_quantizers.227 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.gate_up_proj_weight_quantizers.228 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.gate_up_proj_weight_quantizers.229 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.gate_up_proj_weight_quantizers.230 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.gate_up_proj_weight_quantizers.231 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.gate_up_proj_weight_quantizers.232 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.gate_up_proj_weight_quantizers.233 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.gate_up_proj_weight_quantizers.234 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.28e-01 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.gate_up_proj_weight_quantizers.235 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.gate_up_proj_weight_quantizers.236 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.gate_up_proj_weight_quantizers.237 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.gate_up_proj_weight_quantizers.238 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.gate_up_proj_weight_quantizers.239 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.gate_up_proj_weight_quantizers.240 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.gate_up_proj_weight_quantizers.241 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.18e-01 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.gate_up_proj_weight_quantizers.242 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.gate_up_proj_weight_quantizers.243 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.gate_up_proj_weight_quantizers.244 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.38e-01 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.gate_up_proj_weight_quantizers.245 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.gate_up_proj_weight_quantizers.246 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.gate_up_proj_weight_quantizers.247 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.gate_up_proj_weight_quantizers.248 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.gate_up_proj_weight_quantizers.249 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.gate_up_proj_weight_quantizers.250 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.gate_up_proj_weight_quantizers.251 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.gate_up_proj_weight_quantizers.252 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.gate_up_proj_weight_quantizers.253 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.gate_up_proj_weight_quantizers.254 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.gate_up_proj_weight_quantizers.255 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.down_proj_input_quantizer TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=5.47e+00 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.down_proj_weight_quantizers.0 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.down_proj_weight_quantizers.1 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.down_proj_weight_quantizers.2 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.down_proj_weight_quantizers.3 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.down_proj_weight_quantizers.4 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.down_proj_weight_quantizers.5 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.down_proj_weight_quantizers.6 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.down_proj_weight_quantizers.7 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.down_proj_weight_quantizers.8 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.down_proj_weight_quantizers.9 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.down_proj_weight_quantizers.10 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.down_proj_weight_quantizers.11 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.down_proj_weight_quantizers.12 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.down_proj_weight_quantizers.13 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.99e-01 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.down_proj_weight_quantizers.14 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.15e-01 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.down_proj_weight_quantizers.15 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.23e-01 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.down_proj_weight_quantizers.16 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.down_proj_weight_quantizers.17 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.down_proj_weight_quantizers.18 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.38e-01 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.down_proj_weight_quantizers.19 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.34e-01 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.down_proj_weight_quantizers.20 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.down_proj_weight_quantizers.21 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.down_proj_weight_quantizers.22 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.down_proj_weight_quantizers.23 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.down_proj_weight_quantizers.24 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.down_proj_weight_quantizers.25 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.down_proj_weight_quantizers.26 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.15e-01 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.down_proj_weight_quantizers.27 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.44e-01 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.down_proj_weight_quantizers.28 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.down_proj_weight_quantizers.29 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.down_proj_weight_quantizers.30 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.17e-01 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.down_proj_weight_quantizers.31 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.down_proj_weight_quantizers.32 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.down_proj_weight_quantizers.33 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.18e-01 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.down_proj_weight_quantizers.34 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.13e-01 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.down_proj_weight_quantizers.35 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.down_proj_weight_quantizers.36 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.02e-01 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.down_proj_weight_quantizers.37 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.21e-01 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.down_proj_weight_quantizers.38 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.24e-01 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.down_proj_weight_quantizers.39 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.17e-01 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.down_proj_weight_quantizers.40 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.32e-01 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.down_proj_weight_quantizers.41 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.68e-01 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.down_proj_weight_quantizers.42 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.down_proj_weight_quantizers.43 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.down_proj_weight_quantizers.44 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.down_proj_weight_quantizers.45 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.down_proj_weight_quantizers.46 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.down_proj_weight_quantizers.47 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.down_proj_weight_quantizers.48 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.16e-01 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.down_proj_weight_quantizers.49 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.39e-01 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.down_proj_weight_quantizers.50 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.36e-01 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.down_proj_weight_quantizers.51 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.down_proj_weight_quantizers.52 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.89e-01 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.down_proj_weight_quantizers.53 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.17e-01 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.down_proj_weight_quantizers.54 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.down_proj_weight_quantizers.55 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.down_proj_weight_quantizers.56 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.19e-01 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.down_proj_weight_quantizers.57 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.down_proj_weight_quantizers.58 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.down_proj_weight_quantizers.59 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.49e-01 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.down_proj_weight_quantizers.60 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.down_proj_weight_quantizers.61 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.39e-01 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.down_proj_weight_quantizers.62 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.down_proj_weight_quantizers.63 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.22e-01 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.down_proj_weight_quantizers.64 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.25e-01 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.down_proj_weight_quantizers.65 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.down_proj_weight_quantizers.66 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.down_proj_weight_quantizers.67 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.15e-01 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.down_proj_weight_quantizers.68 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.down_proj_weight_quantizers.69 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.down_proj_weight_quantizers.70 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.62e-01 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.down_proj_weight_quantizers.71 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.66e-01 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.down_proj_weight_quantizers.72 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.down_proj_weight_quantizers.73 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.67e-01 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.down_proj_weight_quantizers.74 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.56e-01 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.down_proj_weight_quantizers.75 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.down_proj_weight_quantizers.76 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.down_proj_weight_quantizers.77 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.down_proj_weight_quantizers.78 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.down_proj_weight_quantizers.79 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.down_proj_weight_quantizers.80 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.down_proj_weight_quantizers.81 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.down_proj_weight_quantizers.82 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.26e-01 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.down_proj_weight_quantizers.83 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.down_proj_weight_quantizers.84 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.down_proj_weight_quantizers.85 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.37e-01 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.down_proj_weight_quantizers.86 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.74e-01 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.down_proj_weight_quantizers.87 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.29e-01 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.down_proj_weight_quantizers.88 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.22e-01 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.down_proj_weight_quantizers.89 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=4.51e-01 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.down_proj_weight_quantizers.90 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.down_proj_weight_quantizers.91 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.down_proj_weight_quantizers.92 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.29e-01 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.down_proj_weight_quantizers.93 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.down_proj_weight_quantizers.94 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.16e-01 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.down_proj_weight_quantizers.95 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.down_proj_weight_quantizers.96 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.40e-01 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.down_proj_weight_quantizers.97 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.33e-01 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.down_proj_weight_quantizers.98 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.down_proj_weight_quantizers.99 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.99e-01 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.down_proj_weight_quantizers.100 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.down_proj_weight_quantizers.101 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.22e-01 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.down_proj_weight_quantizers.102 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.42e-01 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.down_proj_weight_quantizers.103 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.down_proj_weight_quantizers.104 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.down_proj_weight_quantizers.105 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.down_proj_weight_quantizers.106 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.19e-01 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.down_proj_weight_quantizers.107 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.down_proj_weight_quantizers.108 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.down_proj_weight_quantizers.109 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.22e-01 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.down_proj_weight_quantizers.110 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.down_proj_weight_quantizers.111 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.70e-01 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.down_proj_weight_quantizers.112 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.down_proj_weight_quantizers.113 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.down_proj_weight_quantizers.114 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.37e-01 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.down_proj_weight_quantizers.115 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.18e-01 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.down_proj_weight_quantizers.116 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.down_proj_weight_quantizers.117 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.37e-01 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.down_proj_weight_quantizers.118 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.down_proj_weight_quantizers.119 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.35e-01 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.down_proj_weight_quantizers.120 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.down_proj_weight_quantizers.121 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.down_proj_weight_quantizers.122 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.down_proj_weight_quantizers.123 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.down_proj_weight_quantizers.124 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.down_proj_weight_quantizers.125 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.28e-01 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.down_proj_weight_quantizers.126 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.15e-01 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.down_proj_weight_quantizers.127 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.down_proj_weight_quantizers.128 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.27e-01 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.down_proj_weight_quantizers.129 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.down_proj_weight_quantizers.130 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.16e-01 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.down_proj_weight_quantizers.131 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.down_proj_weight_quantizers.132 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.down_proj_weight_quantizers.133 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.down_proj_weight_quantizers.134 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.down_proj_weight_quantizers.135 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.down_proj_weight_quantizers.136 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.down_proj_weight_quantizers.137 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.56e-01 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.down_proj_weight_quantizers.138 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.down_proj_weight_quantizers.139 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.down_proj_weight_quantizers.140 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.down_proj_weight_quantizers.141 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.86e-01 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.down_proj_weight_quantizers.142 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.down_proj_weight_quantizers.143 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.39e-01 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.down_proj_weight_quantizers.144 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.down_proj_weight_quantizers.145 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.down_proj_weight_quantizers.146 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.40e-01 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.down_proj_weight_quantizers.147 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.down_proj_weight_quantizers.148 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.down_proj_weight_quantizers.149 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.down_proj_weight_quantizers.150 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.down_proj_weight_quantizers.151 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.down_proj_weight_quantizers.152 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.down_proj_weight_quantizers.153 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.down_proj_weight_quantizers.154 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.down_proj_weight_quantizers.155 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.56e-01 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.down_proj_weight_quantizers.156 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.down_proj_weight_quantizers.157 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.down_proj_weight_quantizers.158 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.down_proj_weight_quantizers.159 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.40e-01 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.down_proj_weight_quantizers.160 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.down_proj_weight_quantizers.161 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.55e-01 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.down_proj_weight_quantizers.162 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.46e-01 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.down_proj_weight_quantizers.163 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.down_proj_weight_quantizers.164 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.down_proj_weight_quantizers.165 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.down_proj_weight_quantizers.166 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.down_proj_weight_quantizers.167 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.51e-01 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.down_proj_weight_quantizers.168 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.26e-01 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.down_proj_weight_quantizers.169 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.down_proj_weight_quantizers.170 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.down_proj_weight_quantizers.171 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.17e-01 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.down_proj_weight_quantizers.172 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.22e-01 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.down_proj_weight_quantizers.173 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=7.07e-01 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.down_proj_weight_quantizers.174 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.17e-01 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.down_proj_weight_quantizers.175 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.23e-01 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.down_proj_weight_quantizers.176 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.down_proj_weight_quantizers.177 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.54e-01 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.down_proj_weight_quantizers.178 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.down_proj_weight_quantizers.179 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.down_proj_weight_quantizers.180 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.17e-01 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.down_proj_weight_quantizers.181 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.25e-01 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.down_proj_weight_quantizers.182 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.down_proj_weight_quantizers.183 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.down_proj_weight_quantizers.184 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.22e-01 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.down_proj_weight_quantizers.185 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.down_proj_weight_quantizers.186 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.down_proj_weight_quantizers.187 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.10e-01 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.down_proj_weight_quantizers.188 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.31e-01 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.down_proj_weight_quantizers.189 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.down_proj_weight_quantizers.190 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.down_proj_weight_quantizers.191 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.13e-01 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.down_proj_weight_quantizers.192 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.34e-01 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.down_proj_weight_quantizers.193 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.down_proj_weight_quantizers.194 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.down_proj_weight_quantizers.195 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.05e-01 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.down_proj_weight_quantizers.196 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.down_proj_weight_quantizers.197 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.down_proj_weight_quantizers.198 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.29e-01 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.down_proj_weight_quantizers.199 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.down_proj_weight_quantizers.200 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.33e-01 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.down_proj_weight_quantizers.201 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.down_proj_weight_quantizers.202 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.78e-01 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.down_proj_weight_quantizers.203 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.20e-01 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.down_proj_weight_quantizers.204 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.42e-01 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.down_proj_weight_quantizers.205 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.down_proj_weight_quantizers.206 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.13e-01 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.down_proj_weight_quantizers.207 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.down_proj_weight_quantizers.208 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.down_proj_weight_quantizers.209 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.20e-01 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.down_proj_weight_quantizers.210 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.26e-01 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.down_proj_weight_quantizers.211 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.down_proj_weight_quantizers.212 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.down_proj_weight_quantizers.213 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.45e-01 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.down_proj_weight_quantizers.214 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.down_proj_weight_quantizers.215 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.down_proj_weight_quantizers.216 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.down_proj_weight_quantizers.217 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.down_proj_weight_quantizers.218 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.66e-01 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.down_proj_weight_quantizers.219 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.down_proj_weight_quantizers.220 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.down_proj_weight_quantizers.221 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.down_proj_weight_quantizers.222 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.24e-01 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.down_proj_weight_quantizers.223 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.19e-01 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.down_proj_weight_quantizers.224 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.down_proj_weight_quantizers.225 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.down_proj_weight_quantizers.226 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.down_proj_weight_quantizers.227 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.down_proj_weight_quantizers.228 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.down_proj_weight_quantizers.229 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.down_proj_weight_quantizers.230 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.down_proj_weight_quantizers.231 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.down_proj_weight_quantizers.232 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.15e-01 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.down_proj_weight_quantizers.233 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.down_proj_weight_quantizers.234 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.down_proj_weight_quantizers.235 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.26e-01 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.down_proj_weight_quantizers.236 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.down_proj_weight_quantizers.237 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.down_proj_weight_quantizers.238 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.down_proj_weight_quantizers.239 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.down_proj_weight_quantizers.240 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.down_proj_weight_quantizers.241 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.17e-01 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.down_proj_weight_quantizers.242 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.down_proj_weight_quantizers.243 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.23e-01 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.down_proj_weight_quantizers.244 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.13e-01 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.down_proj_weight_quantizers.245 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.22e-01 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.down_proj_weight_quantizers.246 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.17e-01 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.down_proj_weight_quantizers.247 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.53e-01 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.down_proj_weight_quantizers.248 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.40e-01 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.down_proj_weight_quantizers.249 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.28e-01 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.down_proj_weight_quantizers.250 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.down_proj_weight_quantizers.251 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.down_proj_weight_quantizers.252 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.16e-01 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.down_proj_weight_quantizers.253 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.22e-01 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.down_proj_weight_quantizers.254 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.down_proj_weight_quantizers.255 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.10.mlp.shared_experts.gate_proj.input_quantizer TensorQuantizer(disabled) +model.layers.10.mlp.shared_experts.gate_proj.output_quantizer TensorQuantizer(disabled) +model.layers.10.mlp.shared_experts.gate_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.10.mlp.shared_experts.up_proj.input_quantizer TensorQuantizer(disabled) +model.layers.10.mlp.shared_experts.up_proj.output_quantizer TensorQuantizer(disabled) +model.layers.10.mlp.shared_experts.up_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.10.mlp.shared_experts.down_proj.input_quantizer TensorQuantizer(disabled) +model.layers.10.mlp.shared_experts.down_proj.output_quantizer TensorQuantizer(disabled) +model.layers.10.mlp.shared_experts.down_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.11.self_attn.q_a_proj.input_quantizer TensorQuantizer(disabled) +model.layers.11.self_attn.q_a_proj.output_quantizer TensorQuantizer(disabled) +model.layers.11.self_attn.q_a_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.11.self_attn.q_b_proj.input_quantizer TensorQuantizer(disabled) +model.layers.11.self_attn.q_b_proj.output_quantizer TensorQuantizer(disabled) +model.layers.11.self_attn.q_b_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.11.self_attn.kv_a_proj_with_mqa.input_quantizer TensorQuantizer(disabled) +model.layers.11.self_attn.kv_a_proj_with_mqa.output_quantizer TensorQuantizer(disabled) +model.layers.11.self_attn.kv_a_proj_with_mqa.weight_quantizer TensorQuantizer(disabled) +model.layers.11.self_attn.kv_b_proj.input_quantizer TensorQuantizer(disabled) +model.layers.11.self_attn.kv_b_proj.output_quantizer TensorQuantizer(disabled) +model.layers.11.self_attn.kv_b_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.11.self_attn.o_proj.input_quantizer TensorQuantizer(disabled) +model.layers.11.self_attn.o_proj.output_quantizer TensorQuantizer(disabled) +model.layers.11.self_attn.o_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.11.self_attn.q_bmm_quantizer TensorQuantizer(disabled) +model.layers.11.self_attn.k_bmm_quantizer TensorQuantizer((4, 3) bit fake per-tensor amax=4.48e+02(const) calibrator=MaxCalibrator quant) +model.layers.11.self_attn.v_bmm_quantizer TensorQuantizer((4, 3) bit fake per-tensor amax=4.48e+02(const) calibrator=MaxCalibrator quant) +model.layers.11.self_attn.p_bmm_quantizer TensorQuantizer(disabled) +model.layers.11.mlp.experts.gate_up_proj_input_quantizer TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=5.35e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.gate_up_proj_weight_quantizers.0 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.gate_up_proj_weight_quantizers.1 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.67e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.gate_up_proj_weight_quantizers.2 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.gate_up_proj_weight_quantizers.3 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.gate_up_proj_weight_quantizers.4 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.23e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.gate_up_proj_weight_quantizers.5 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.gate_up_proj_weight_quantizers.6 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.gate_up_proj_weight_quantizers.7 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.gate_up_proj_weight_quantizers.8 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.gate_up_proj_weight_quantizers.9 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.gate_up_proj_weight_quantizers.10 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.gate_up_proj_weight_quantizers.11 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.gate_up_proj_weight_quantizers.12 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.gate_up_proj_weight_quantizers.13 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.gate_up_proj_weight_quantizers.14 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.gate_up_proj_weight_quantizers.15 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.gate_up_proj_weight_quantizers.16 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.26e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.gate_up_proj_weight_quantizers.17 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.gate_up_proj_weight_quantizers.18 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.26e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.gate_up_proj_weight_quantizers.19 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.14e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.gate_up_proj_weight_quantizers.20 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.gate_up_proj_weight_quantizers.21 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.gate_up_proj_weight_quantizers.22 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.gate_up_proj_weight_quantizers.23 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.gate_up_proj_weight_quantizers.24 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.gate_up_proj_weight_quantizers.25 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.gate_up_proj_weight_quantizers.26 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.gate_up_proj_weight_quantizers.27 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.gate_up_proj_weight_quantizers.28 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.37e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.gate_up_proj_weight_quantizers.29 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.gate_up_proj_weight_quantizers.30 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.55e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.gate_up_proj_weight_quantizers.31 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.gate_up_proj_weight_quantizers.32 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.gate_up_proj_weight_quantizers.33 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.gate_up_proj_weight_quantizers.34 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.48e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.gate_up_proj_weight_quantizers.35 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.gate_up_proj_weight_quantizers.36 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.gate_up_proj_weight_quantizers.37 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.gate_up_proj_weight_quantizers.38 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.gate_up_proj_weight_quantizers.39 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.gate_up_proj_weight_quantizers.40 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.gate_up_proj_weight_quantizers.41 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.gate_up_proj_weight_quantizers.42 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.gate_up_proj_weight_quantizers.43 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.gate_up_proj_weight_quantizers.44 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.gate_up_proj_weight_quantizers.45 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.82e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.gate_up_proj_weight_quantizers.46 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.gate_up_proj_weight_quantizers.47 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.gate_up_proj_weight_quantizers.48 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.gate_up_proj_weight_quantizers.49 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.gate_up_proj_weight_quantizers.50 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.gate_up_proj_weight_quantizers.51 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.gate_up_proj_weight_quantizers.52 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.gate_up_proj_weight_quantizers.53 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.26e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.gate_up_proj_weight_quantizers.54 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.15e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.gate_up_proj_weight_quantizers.55 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.gate_up_proj_weight_quantizers.56 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.gate_up_proj_weight_quantizers.57 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.gate_up_proj_weight_quantizers.58 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.gate_up_proj_weight_quantizers.59 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.gate_up_proj_weight_quantizers.60 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.gate_up_proj_weight_quantizers.61 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.gate_up_proj_weight_quantizers.62 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.gate_up_proj_weight_quantizers.63 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.gate_up_proj_weight_quantizers.64 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.14e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.gate_up_proj_weight_quantizers.65 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.gate_up_proj_weight_quantizers.66 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.gate_up_proj_weight_quantizers.67 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.gate_up_proj_weight_quantizers.68 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.23e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.gate_up_proj_weight_quantizers.69 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.33e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.gate_up_proj_weight_quantizers.70 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.gate_up_proj_weight_quantizers.71 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.gate_up_proj_weight_quantizers.72 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.gate_up_proj_weight_quantizers.73 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.gate_up_proj_weight_quantizers.74 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.gate_up_proj_weight_quantizers.75 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.14e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.gate_up_proj_weight_quantizers.76 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.gate_up_proj_weight_quantizers.77 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.97e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.gate_up_proj_weight_quantizers.78 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.gate_up_proj_weight_quantizers.79 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.gate_up_proj_weight_quantizers.80 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.96e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.gate_up_proj_weight_quantizers.81 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.gate_up_proj_weight_quantizers.82 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.gate_up_proj_weight_quantizers.83 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.32e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.gate_up_proj_weight_quantizers.84 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.gate_up_proj_weight_quantizers.85 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.gate_up_proj_weight_quantizers.86 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.26e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.gate_up_proj_weight_quantizers.87 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.gate_up_proj_weight_quantizers.88 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.gate_up_proj_weight_quantizers.89 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.gate_up_proj_weight_quantizers.90 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.gate_up_proj_weight_quantizers.91 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.gate_up_proj_weight_quantizers.92 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.86e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.gate_up_proj_weight_quantizers.93 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.gate_up_proj_weight_quantizers.94 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.gate_up_proj_weight_quantizers.95 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.gate_up_proj_weight_quantizers.96 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.gate_up_proj_weight_quantizers.97 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.gate_up_proj_weight_quantizers.98 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.gate_up_proj_weight_quantizers.99 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.gate_up_proj_weight_quantizers.100 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.gate_up_proj_weight_quantizers.101 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.gate_up_proj_weight_quantizers.102 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.gate_up_proj_weight_quantizers.103 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.gate_up_proj_weight_quantizers.104 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.gate_up_proj_weight_quantizers.105 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.13e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.gate_up_proj_weight_quantizers.106 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.gate_up_proj_weight_quantizers.107 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.gate_up_proj_weight_quantizers.108 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.gate_up_proj_weight_quantizers.109 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.gate_up_proj_weight_quantizers.110 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.gate_up_proj_weight_quantizers.111 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.48e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.gate_up_proj_weight_quantizers.112 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.gate_up_proj_weight_quantizers.113 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.gate_up_proj_weight_quantizers.114 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.81e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.gate_up_proj_weight_quantizers.115 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.gate_up_proj_weight_quantizers.116 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.gate_up_proj_weight_quantizers.117 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.gate_up_proj_weight_quantizers.118 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.gate_up_proj_weight_quantizers.119 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.gate_up_proj_weight_quantizers.120 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.gate_up_proj_weight_quantizers.121 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.gate_up_proj_weight_quantizers.122 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.gate_up_proj_weight_quantizers.123 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.gate_up_proj_weight_quantizers.124 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.gate_up_proj_weight_quantizers.125 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.gate_up_proj_weight_quantizers.126 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.41e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.gate_up_proj_weight_quantizers.127 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.gate_up_proj_weight_quantizers.128 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.gate_up_proj_weight_quantizers.129 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.gate_up_proj_weight_quantizers.130 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.46e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.gate_up_proj_weight_quantizers.131 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.41e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.gate_up_proj_weight_quantizers.132 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.gate_up_proj_weight_quantizers.133 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.gate_up_proj_weight_quantizers.134 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.gate_up_proj_weight_quantizers.135 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.gate_up_proj_weight_quantizers.136 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.20e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.gate_up_proj_weight_quantizers.137 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.gate_up_proj_weight_quantizers.138 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.gate_up_proj_weight_quantizers.139 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.gate_up_proj_weight_quantizers.140 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.gate_up_proj_weight_quantizers.141 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.gate_up_proj_weight_quantizers.142 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.gate_up_proj_weight_quantizers.143 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.gate_up_proj_weight_quantizers.144 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.gate_up_proj_weight_quantizers.145 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.gate_up_proj_weight_quantizers.146 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.gate_up_proj_weight_quantizers.147 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.gate_up_proj_weight_quantizers.148 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.gate_up_proj_weight_quantizers.149 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.gate_up_proj_weight_quantizers.150 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.gate_up_proj_weight_quantizers.151 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.gate_up_proj_weight_quantizers.152 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.gate_up_proj_weight_quantizers.153 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.gate_up_proj_weight_quantizers.154 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.99e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.gate_up_proj_weight_quantizers.155 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.gate_up_proj_weight_quantizers.156 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.47e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.gate_up_proj_weight_quantizers.157 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.80e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.gate_up_proj_weight_quantizers.158 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.22e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.gate_up_proj_weight_quantizers.159 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.gate_up_proj_weight_quantizers.160 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.18e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.gate_up_proj_weight_quantizers.161 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.gate_up_proj_weight_quantizers.162 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.gate_up_proj_weight_quantizers.163 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.30e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.gate_up_proj_weight_quantizers.164 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.gate_up_proj_weight_quantizers.165 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.gate_up_proj_weight_quantizers.166 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.gate_up_proj_weight_quantizers.167 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.gate_up_proj_weight_quantizers.168 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.gate_up_proj_weight_quantizers.169 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.gate_up_proj_weight_quantizers.170 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.77e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.gate_up_proj_weight_quantizers.171 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.gate_up_proj_weight_quantizers.172 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.gate_up_proj_weight_quantizers.173 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.67e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.gate_up_proj_weight_quantizers.174 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.gate_up_proj_weight_quantizers.175 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.14e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.gate_up_proj_weight_quantizers.176 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.gate_up_proj_weight_quantizers.177 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.gate_up_proj_weight_quantizers.178 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.gate_up_proj_weight_quantizers.179 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.93e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.gate_up_proj_weight_quantizers.180 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.gate_up_proj_weight_quantizers.181 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.19e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.gate_up_proj_weight_quantizers.182 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.72e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.gate_up_proj_weight_quantizers.183 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.gate_up_proj_weight_quantizers.184 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.gate_up_proj_weight_quantizers.185 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.gate_up_proj_weight_quantizers.186 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.gate_up_proj_weight_quantizers.187 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.gate_up_proj_weight_quantizers.188 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.gate_up_proj_weight_quantizers.189 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.gate_up_proj_weight_quantizers.190 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.gate_up_proj_weight_quantizers.191 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.gate_up_proj_weight_quantizers.192 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.21e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.gate_up_proj_weight_quantizers.193 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.gate_up_proj_weight_quantizers.194 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.gate_up_proj_weight_quantizers.195 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.gate_up_proj_weight_quantizers.196 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.gate_up_proj_weight_quantizers.197 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.37e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.gate_up_proj_weight_quantizers.198 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.gate_up_proj_weight_quantizers.199 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.gate_up_proj_weight_quantizers.200 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.gate_up_proj_weight_quantizers.201 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.gate_up_proj_weight_quantizers.202 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.46e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.gate_up_proj_weight_quantizers.203 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.gate_up_proj_weight_quantizers.204 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.20e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.gate_up_proj_weight_quantizers.205 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.gate_up_proj_weight_quantizers.206 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.gate_up_proj_weight_quantizers.207 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.47e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.gate_up_proj_weight_quantizers.208 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.gate_up_proj_weight_quantizers.209 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.gate_up_proj_weight_quantizers.210 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.19e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.gate_up_proj_weight_quantizers.211 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.gate_up_proj_weight_quantizers.212 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.17e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.gate_up_proj_weight_quantizers.213 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.16e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.gate_up_proj_weight_quantizers.214 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.gate_up_proj_weight_quantizers.215 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.gate_up_proj_weight_quantizers.216 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.gate_up_proj_weight_quantizers.217 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.gate_up_proj_weight_quantizers.218 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.gate_up_proj_weight_quantizers.219 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.gate_up_proj_weight_quantizers.220 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.21e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.gate_up_proj_weight_quantizers.221 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.gate_up_proj_weight_quantizers.222 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.gate_up_proj_weight_quantizers.223 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.gate_up_proj_weight_quantizers.224 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.14e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.gate_up_proj_weight_quantizers.225 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.gate_up_proj_weight_quantizers.226 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.gate_up_proj_weight_quantizers.227 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.gate_up_proj_weight_quantizers.228 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.gate_up_proj_weight_quantizers.229 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.gate_up_proj_weight_quantizers.230 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.55e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.gate_up_proj_weight_quantizers.231 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.15e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.gate_up_proj_weight_quantizers.232 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.gate_up_proj_weight_quantizers.233 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.gate_up_proj_weight_quantizers.234 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.gate_up_proj_weight_quantizers.235 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.59e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.gate_up_proj_weight_quantizers.236 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.96e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.gate_up_proj_weight_quantizers.237 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.gate_up_proj_weight_quantizers.238 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.gate_up_proj_weight_quantizers.239 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.gate_up_proj_weight_quantizers.240 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.gate_up_proj_weight_quantizers.241 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.gate_up_proj_weight_quantizers.242 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.gate_up_proj_weight_quantizers.243 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.15e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.gate_up_proj_weight_quantizers.244 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.gate_up_proj_weight_quantizers.245 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.gate_up_proj_weight_quantizers.246 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.gate_up_proj_weight_quantizers.247 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.gate_up_proj_weight_quantizers.248 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.gate_up_proj_weight_quantizers.249 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.gate_up_proj_weight_quantizers.250 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.gate_up_proj_weight_quantizers.251 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.gate_up_proj_weight_quantizers.252 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.gate_up_proj_weight_quantizers.253 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.60e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.gate_up_proj_weight_quantizers.254 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.gate_up_proj_weight_quantizers.255 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.down_proj_input_quantizer TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.28e+00 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.down_proj_weight_quantizers.0 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.47e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.down_proj_weight_quantizers.1 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.13e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.down_proj_weight_quantizers.2 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.40e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.down_proj_weight_quantizers.3 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.down_proj_weight_quantizers.4 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.down_proj_weight_quantizers.5 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.35e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.down_proj_weight_quantizers.6 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.down_proj_weight_quantizers.7 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.down_proj_weight_quantizers.8 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.21e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.down_proj_weight_quantizers.9 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.40e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.down_proj_weight_quantizers.10 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.down_proj_weight_quantizers.11 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.down_proj_weight_quantizers.12 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.down_proj_weight_quantizers.13 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.25e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.down_proj_weight_quantizers.14 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.down_proj_weight_quantizers.15 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.05e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.down_proj_weight_quantizers.16 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.49e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.down_proj_weight_quantizers.17 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.39e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.down_proj_weight_quantizers.18 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.32e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.down_proj_weight_quantizers.19 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.18e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.down_proj_weight_quantizers.20 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.down_proj_weight_quantizers.21 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.down_proj_weight_quantizers.22 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.19e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.down_proj_weight_quantizers.23 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.down_proj_weight_quantizers.24 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.down_proj_weight_quantizers.25 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.31e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.down_proj_weight_quantizers.26 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.27e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.down_proj_weight_quantizers.27 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.27e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.down_proj_weight_quantizers.28 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.26e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.down_proj_weight_quantizers.29 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.down_proj_weight_quantizers.30 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.down_proj_weight_quantizers.31 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.99e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.down_proj_weight_quantizers.32 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.28e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.down_proj_weight_quantizers.33 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.down_proj_weight_quantizers.34 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.89e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.down_proj_weight_quantizers.35 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.26e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.down_proj_weight_quantizers.36 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.down_proj_weight_quantizers.37 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.37e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.down_proj_weight_quantizers.38 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.19e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.down_proj_weight_quantizers.39 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.down_proj_weight_quantizers.40 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.down_proj_weight_quantizers.41 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.45e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.down_proj_weight_quantizers.42 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.22e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.down_proj_weight_quantizers.43 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.34e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.down_proj_weight_quantizers.44 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.21e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.down_proj_weight_quantizers.45 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.down_proj_weight_quantizers.46 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.down_proj_weight_quantizers.47 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.down_proj_weight_quantizers.48 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.down_proj_weight_quantizers.49 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.down_proj_weight_quantizers.50 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.down_proj_weight_quantizers.51 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.13e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.down_proj_weight_quantizers.52 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.down_proj_weight_quantizers.53 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.28e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.down_proj_weight_quantizers.54 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.72e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.down_proj_weight_quantizers.55 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.46e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.down_proj_weight_quantizers.56 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.down_proj_weight_quantizers.57 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.down_proj_weight_quantizers.58 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.29e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.down_proj_weight_quantizers.59 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.19e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.down_proj_weight_quantizers.60 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.down_proj_weight_quantizers.61 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.39e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.down_proj_weight_quantizers.62 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.down_proj_weight_quantizers.63 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.74e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.down_proj_weight_quantizers.64 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.21e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.down_proj_weight_quantizers.65 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.down_proj_weight_quantizers.66 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.down_proj_weight_quantizers.67 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.28e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.down_proj_weight_quantizers.68 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.14e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.down_proj_weight_quantizers.69 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.down_proj_weight_quantizers.70 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.18e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.down_proj_weight_quantizers.71 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.down_proj_weight_quantizers.72 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.down_proj_weight_quantizers.73 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.down_proj_weight_quantizers.74 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.down_proj_weight_quantizers.75 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.44e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.down_proj_weight_quantizers.76 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.down_proj_weight_quantizers.77 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.down_proj_weight_quantizers.78 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.down_proj_weight_quantizers.79 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.76e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.down_proj_weight_quantizers.80 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.down_proj_weight_quantizers.81 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.25e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.down_proj_weight_quantizers.82 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.down_proj_weight_quantizers.83 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.57e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.down_proj_weight_quantizers.84 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.down_proj_weight_quantizers.85 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.down_proj_weight_quantizers.86 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.down_proj_weight_quantizers.87 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.down_proj_weight_quantizers.88 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.24e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.down_proj_weight_quantizers.89 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.down_proj_weight_quantizers.90 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.down_proj_weight_quantizers.91 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.down_proj_weight_quantizers.92 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.55e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.down_proj_weight_quantizers.93 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.down_proj_weight_quantizers.94 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.down_proj_weight_quantizers.95 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.down_proj_weight_quantizers.96 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.down_proj_weight_quantizers.97 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.18e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.down_proj_weight_quantizers.98 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.down_proj_weight_quantizers.99 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.down_proj_weight_quantizers.100 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.46e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.down_proj_weight_quantizers.101 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.26e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.down_proj_weight_quantizers.102 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.down_proj_weight_quantizers.103 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.down_proj_weight_quantizers.104 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.17e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.down_proj_weight_quantizers.105 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.26e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.down_proj_weight_quantizers.106 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.39e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.down_proj_weight_quantizers.107 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.down_proj_weight_quantizers.108 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.53e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.down_proj_weight_quantizers.109 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.16e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.down_proj_weight_quantizers.110 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.52e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.down_proj_weight_quantizers.111 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.83e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.down_proj_weight_quantizers.112 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.32e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.down_proj_weight_quantizers.113 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.down_proj_weight_quantizers.114 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.52e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.down_proj_weight_quantizers.115 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.17e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.down_proj_weight_quantizers.116 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.down_proj_weight_quantizers.117 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.33e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.down_proj_weight_quantizers.118 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.down_proj_weight_quantizers.119 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.down_proj_weight_quantizers.120 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.down_proj_weight_quantizers.121 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.down_proj_weight_quantizers.122 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.37e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.down_proj_weight_quantizers.123 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.down_proj_weight_quantizers.124 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.down_proj_weight_quantizers.125 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.17e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.down_proj_weight_quantizers.126 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.46e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.down_proj_weight_quantizers.127 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.down_proj_weight_quantizers.128 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.down_proj_weight_quantizers.129 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.down_proj_weight_quantizers.130 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.66e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.down_proj_weight_quantizers.131 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.80e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.down_proj_weight_quantizers.132 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.down_proj_weight_quantizers.133 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.down_proj_weight_quantizers.134 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.down_proj_weight_quantizers.135 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.27e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.down_proj_weight_quantizers.136 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.47e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.down_proj_weight_quantizers.137 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.24e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.down_proj_weight_quantizers.138 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.14e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.down_proj_weight_quantizers.139 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.28e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.down_proj_weight_quantizers.140 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.39e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.down_proj_weight_quantizers.141 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.68e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.down_proj_weight_quantizers.142 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.30e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.down_proj_weight_quantizers.143 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.down_proj_weight_quantizers.144 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.13e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.down_proj_weight_quantizers.145 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.26e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.down_proj_weight_quantizers.146 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.17e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.down_proj_weight_quantizers.147 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.down_proj_weight_quantizers.148 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.75e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.down_proj_weight_quantizers.149 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.22e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.down_proj_weight_quantizers.150 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.20e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.down_proj_weight_quantizers.151 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.down_proj_weight_quantizers.152 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.down_proj_weight_quantizers.153 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.down_proj_weight_quantizers.154 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.52e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.down_proj_weight_quantizers.155 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.38e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.down_proj_weight_quantizers.156 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.down_proj_weight_quantizers.157 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.down_proj_weight_quantizers.158 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.37e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.down_proj_weight_quantizers.159 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.down_proj_weight_quantizers.160 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.57e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.down_proj_weight_quantizers.161 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.down_proj_weight_quantizers.162 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.down_proj_weight_quantizers.163 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.14e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.down_proj_weight_quantizers.164 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.down_proj_weight_quantizers.165 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.23e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.down_proj_weight_quantizers.166 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.down_proj_weight_quantizers.167 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.down_proj_weight_quantizers.168 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.down_proj_weight_quantizers.169 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.57e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.down_proj_weight_quantizers.170 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.24e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.down_proj_weight_quantizers.171 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.down_proj_weight_quantizers.172 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.down_proj_weight_quantizers.173 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.54e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.down_proj_weight_quantizers.174 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.24e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.down_proj_weight_quantizers.175 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.67e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.down_proj_weight_quantizers.176 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.56e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.down_proj_weight_quantizers.177 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.25e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.down_proj_weight_quantizers.178 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.32e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.down_proj_weight_quantizers.179 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.25e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.down_proj_weight_quantizers.180 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.49e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.down_proj_weight_quantizers.181 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.27e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.down_proj_weight_quantizers.182 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.85e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.down_proj_weight_quantizers.183 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.down_proj_weight_quantizers.184 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.39e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.down_proj_weight_quantizers.185 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.18e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.down_proj_weight_quantizers.186 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.43e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.down_proj_weight_quantizers.187 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.14e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.down_proj_weight_quantizers.188 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.down_proj_weight_quantizers.189 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.down_proj_weight_quantizers.190 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.down_proj_weight_quantizers.191 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.64e-02 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.down_proj_weight_quantizers.192 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.13e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.down_proj_weight_quantizers.193 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.down_proj_weight_quantizers.194 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.down_proj_weight_quantizers.195 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.down_proj_weight_quantizers.196 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.down_proj_weight_quantizers.197 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.down_proj_weight_quantizers.198 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.down_proj_weight_quantizers.199 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.76e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.down_proj_weight_quantizers.200 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.down_proj_weight_quantizers.201 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.35e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.down_proj_weight_quantizers.202 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.77e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.down_proj_weight_quantizers.203 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.down_proj_weight_quantizers.204 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.14e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.down_proj_weight_quantizers.205 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.down_proj_weight_quantizers.206 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.down_proj_weight_quantizers.207 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.07e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.down_proj_weight_quantizers.208 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.down_proj_weight_quantizers.209 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.down_proj_weight_quantizers.210 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.90e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.down_proj_weight_quantizers.211 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.53e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.down_proj_weight_quantizers.212 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.down_proj_weight_quantizers.213 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.16e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.down_proj_weight_quantizers.214 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.15e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.down_proj_weight_quantizers.215 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.down_proj_weight_quantizers.216 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.90e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.down_proj_weight_quantizers.217 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.down_proj_weight_quantizers.218 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.down_proj_weight_quantizers.219 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.down_proj_weight_quantizers.220 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.19e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.down_proj_weight_quantizers.221 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.46e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.down_proj_weight_quantizers.222 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.64e-02 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.down_proj_weight_quantizers.223 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.down_proj_weight_quantizers.224 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.16e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.down_proj_weight_quantizers.225 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.down_proj_weight_quantizers.226 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.down_proj_weight_quantizers.227 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.down_proj_weight_quantizers.228 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.69e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.down_proj_weight_quantizers.229 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.down_proj_weight_quantizers.230 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.down_proj_weight_quantizers.231 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.56e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.down_proj_weight_quantizers.232 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.down_proj_weight_quantizers.233 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.down_proj_weight_quantizers.234 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.38e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.down_proj_weight_quantizers.235 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.37e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.down_proj_weight_quantizers.236 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.40e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.down_proj_weight_quantizers.237 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.down_proj_weight_quantizers.238 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.down_proj_weight_quantizers.239 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.down_proj_weight_quantizers.240 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.down_proj_weight_quantizers.241 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.53e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.down_proj_weight_quantizers.242 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.49e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.down_proj_weight_quantizers.243 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.21e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.down_proj_weight_quantizers.244 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.21e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.down_proj_weight_quantizers.245 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.down_proj_weight_quantizers.246 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.down_proj_weight_quantizers.247 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.down_proj_weight_quantizers.248 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.down_proj_weight_quantizers.249 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.39e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.down_proj_weight_quantizers.250 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.down_proj_weight_quantizers.251 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.40e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.down_proj_weight_quantizers.252 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.down_proj_weight_quantizers.253 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.down_proj_weight_quantizers.254 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.18e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.down_proj_weight_quantizers.255 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.47e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.shared_experts.gate_proj.input_quantizer TensorQuantizer(disabled) +model.layers.11.mlp.shared_experts.gate_proj.output_quantizer TensorQuantizer(disabled) +model.layers.11.mlp.shared_experts.gate_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.11.mlp.shared_experts.up_proj.input_quantizer TensorQuantizer(disabled) +model.layers.11.mlp.shared_experts.up_proj.output_quantizer TensorQuantizer(disabled) +model.layers.11.mlp.shared_experts.up_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.11.mlp.shared_experts.down_proj.input_quantizer TensorQuantizer(disabled) +model.layers.11.mlp.shared_experts.down_proj.output_quantizer TensorQuantizer(disabled) +model.layers.11.mlp.shared_experts.down_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.12.self_attn.q_a_proj.input_quantizer TensorQuantizer(disabled) +model.layers.12.self_attn.q_a_proj.output_quantizer TensorQuantizer(disabled) +model.layers.12.self_attn.q_a_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.12.self_attn.q_b_proj.input_quantizer TensorQuantizer(disabled) +model.layers.12.self_attn.q_b_proj.output_quantizer TensorQuantizer(disabled) +model.layers.12.self_attn.q_b_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.12.self_attn.kv_a_proj_with_mqa.input_quantizer TensorQuantizer(disabled) +model.layers.12.self_attn.kv_a_proj_with_mqa.output_quantizer TensorQuantizer(disabled) +model.layers.12.self_attn.kv_a_proj_with_mqa.weight_quantizer TensorQuantizer(disabled) +model.layers.12.self_attn.kv_b_proj.input_quantizer TensorQuantizer(disabled) +model.layers.12.self_attn.kv_b_proj.output_quantizer TensorQuantizer(disabled) +model.layers.12.self_attn.kv_b_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.12.self_attn.o_proj.input_quantizer TensorQuantizer(disabled) +model.layers.12.self_attn.o_proj.output_quantizer TensorQuantizer(disabled) +model.layers.12.self_attn.o_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.12.self_attn.q_bmm_quantizer TensorQuantizer(disabled) +model.layers.12.self_attn.k_bmm_quantizer TensorQuantizer((4, 3) bit fake per-tensor amax=4.48e+02(const) calibrator=MaxCalibrator quant) +model.layers.12.self_attn.v_bmm_quantizer TensorQuantizer((4, 3) bit fake per-tensor amax=4.48e+02(const) calibrator=MaxCalibrator quant) +model.layers.12.self_attn.p_bmm_quantizer TensorQuantizer(disabled) +model.layers.12.mlp.experts.gate_up_proj_input_quantizer TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=4.80e-01 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.gate_up_proj_weight_quantizers.0 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.88e-01 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.gate_up_proj_weight_quantizers.1 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.gate_up_proj_weight_quantizers.2 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.77e-01 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.gate_up_proj_weight_quantizers.3 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.gate_up_proj_weight_quantizers.4 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.gate_up_proj_weight_quantizers.5 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.gate_up_proj_weight_quantizers.6 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.gate_up_proj_weight_quantizers.7 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.59e-01 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.gate_up_proj_weight_quantizers.8 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.gate_up_proj_weight_quantizers.9 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.gate_up_proj_weight_quantizers.10 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.15e-01 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.gate_up_proj_weight_quantizers.11 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.18e-01 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.gate_up_proj_weight_quantizers.12 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.gate_up_proj_weight_quantizers.13 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.gate_up_proj_weight_quantizers.14 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.gate_up_proj_weight_quantizers.15 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.gate_up_proj_weight_quantizers.16 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.gate_up_proj_weight_quantizers.17 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.87e-01 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.gate_up_proj_weight_quantizers.18 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.gate_up_proj_weight_quantizers.19 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.gate_up_proj_weight_quantizers.20 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.gate_up_proj_weight_quantizers.21 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.gate_up_proj_weight_quantizers.22 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.53e-01 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.gate_up_proj_weight_quantizers.23 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.gate_up_proj_weight_quantizers.24 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.gate_up_proj_weight_quantizers.25 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.gate_up_proj_weight_quantizers.26 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.gate_up_proj_weight_quantizers.27 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.gate_up_proj_weight_quantizers.28 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.gate_up_proj_weight_quantizers.29 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.gate_up_proj_weight_quantizers.30 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.gate_up_proj_weight_quantizers.31 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.18e-01 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.gate_up_proj_weight_quantizers.32 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.30e-01 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.gate_up_proj_weight_quantizers.33 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.gate_up_proj_weight_quantizers.34 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.gate_up_proj_weight_quantizers.35 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.gate_up_proj_weight_quantizers.36 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.74e-01 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.gate_up_proj_weight_quantizers.37 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.gate_up_proj_weight_quantizers.38 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.gate_up_proj_weight_quantizers.39 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.gate_up_proj_weight_quantizers.40 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.05e-01 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.gate_up_proj_weight_quantizers.41 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.gate_up_proj_weight_quantizers.42 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.gate_up_proj_weight_quantizers.43 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.gate_up_proj_weight_quantizers.44 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.gate_up_proj_weight_quantizers.45 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.gate_up_proj_weight_quantizers.46 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.gate_up_proj_weight_quantizers.47 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.20e-01 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.gate_up_proj_weight_quantizers.48 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.gate_up_proj_weight_quantizers.49 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.18e-01 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.gate_up_proj_weight_quantizers.50 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.gate_up_proj_weight_quantizers.51 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.gate_up_proj_weight_quantizers.52 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.gate_up_proj_weight_quantizers.53 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.38e-01 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.gate_up_proj_weight_quantizers.54 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.gate_up_proj_weight_quantizers.55 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.gate_up_proj_weight_quantizers.56 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.gate_up_proj_weight_quantizers.57 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.68e-01 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.gate_up_proj_weight_quantizers.58 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.52e-01 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.gate_up_proj_weight_quantizers.59 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.63e-01 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.gate_up_proj_weight_quantizers.60 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.gate_up_proj_weight_quantizers.61 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.75e-01 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.gate_up_proj_weight_quantizers.62 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.gate_up_proj_weight_quantizers.63 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.05e-01 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.gate_up_proj_weight_quantizers.64 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.gate_up_proj_weight_quantizers.65 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.gate_up_proj_weight_quantizers.66 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.76e-01 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.gate_up_proj_weight_quantizers.67 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.39e-01 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.gate_up_proj_weight_quantizers.68 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.39e-01 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.gate_up_proj_weight_quantizers.69 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.gate_up_proj_weight_quantizers.70 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.gate_up_proj_weight_quantizers.71 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.gate_up_proj_weight_quantizers.72 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.gate_up_proj_weight_quantizers.73 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.20e-01 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.gate_up_proj_weight_quantizers.74 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.gate_up_proj_weight_quantizers.75 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.gate_up_proj_weight_quantizers.76 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.gate_up_proj_weight_quantizers.77 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.59e-01 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.gate_up_proj_weight_quantizers.78 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.46e-01 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.gate_up_proj_weight_quantizers.79 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.55e-01 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.gate_up_proj_weight_quantizers.80 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.gate_up_proj_weight_quantizers.81 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.gate_up_proj_weight_quantizers.82 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.gate_up_proj_weight_quantizers.83 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.gate_up_proj_weight_quantizers.84 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=5.08e-01 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.gate_up_proj_weight_quantizers.85 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.gate_up_proj_weight_quantizers.86 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.gate_up_proj_weight_quantizers.87 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.14e-01 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.gate_up_proj_weight_quantizers.88 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.gate_up_proj_weight_quantizers.89 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.gate_up_proj_weight_quantizers.90 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.gate_up_proj_weight_quantizers.91 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.gate_up_proj_weight_quantizers.92 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.19e-01 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.gate_up_proj_weight_quantizers.93 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.gate_up_proj_weight_quantizers.94 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.gate_up_proj_weight_quantizers.95 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.58e-01 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.gate_up_proj_weight_quantizers.96 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.gate_up_proj_weight_quantizers.97 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.gate_up_proj_weight_quantizers.98 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.47e-01 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.gate_up_proj_weight_quantizers.99 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.43e-01 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.gate_up_proj_weight_quantizers.100 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.gate_up_proj_weight_quantizers.101 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.42e-01 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.gate_up_proj_weight_quantizers.102 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.gate_up_proj_weight_quantizers.103 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.gate_up_proj_weight_quantizers.104 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.gate_up_proj_weight_quantizers.105 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.gate_up_proj_weight_quantizers.106 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.gate_up_proj_weight_quantizers.107 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.gate_up_proj_weight_quantizers.108 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.04e-01 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.gate_up_proj_weight_quantizers.109 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.gate_up_proj_weight_quantizers.110 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.13e-01 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.gate_up_proj_weight_quantizers.111 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.gate_up_proj_weight_quantizers.112 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.97e-01 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.gate_up_proj_weight_quantizers.113 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.gate_up_proj_weight_quantizers.114 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.12e-01 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.gate_up_proj_weight_quantizers.115 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.20e-01 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.gate_up_proj_weight_quantizers.116 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.29e-01 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.gate_up_proj_weight_quantizers.117 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.23e-01 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.gate_up_proj_weight_quantizers.118 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.15e-01 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.gate_up_proj_weight_quantizers.119 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.34e-01 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.gate_up_proj_weight_quantizers.120 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.76e-01 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.gate_up_proj_weight_quantizers.121 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.gate_up_proj_weight_quantizers.122 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.gate_up_proj_weight_quantizers.123 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.gate_up_proj_weight_quantizers.124 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.gate_up_proj_weight_quantizers.125 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.gate_up_proj_weight_quantizers.126 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.gate_up_proj_weight_quantizers.127 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.gate_up_proj_weight_quantizers.128 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.33e-01 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.gate_up_proj_weight_quantizers.129 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.gate_up_proj_weight_quantizers.130 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.gate_up_proj_weight_quantizers.131 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.gate_up_proj_weight_quantizers.132 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.gate_up_proj_weight_quantizers.133 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.17e-01 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.gate_up_proj_weight_quantizers.134 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.66e-01 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.gate_up_proj_weight_quantizers.135 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.21e-01 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.gate_up_proj_weight_quantizers.136 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.gate_up_proj_weight_quantizers.137 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.gate_up_proj_weight_quantizers.138 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.gate_up_proj_weight_quantizers.139 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.gate_up_proj_weight_quantizers.140 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.44e-01 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.gate_up_proj_weight_quantizers.141 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.gate_up_proj_weight_quantizers.142 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.gate_up_proj_weight_quantizers.143 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.61e-01 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.gate_up_proj_weight_quantizers.144 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.gate_up_proj_weight_quantizers.145 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.16e-01 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.gate_up_proj_weight_quantizers.146 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.16e-01 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.gate_up_proj_weight_quantizers.147 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.gate_up_proj_weight_quantizers.148 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.gate_up_proj_weight_quantizers.149 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.gate_up_proj_weight_quantizers.150 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.gate_up_proj_weight_quantizers.151 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.gate_up_proj_weight_quantizers.152 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.22e-01 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.gate_up_proj_weight_quantizers.153 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.gate_up_proj_weight_quantizers.154 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.gate_up_proj_weight_quantizers.155 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.51e-01 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.gate_up_proj_weight_quantizers.156 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.gate_up_proj_weight_quantizers.157 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.13e-01 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.gate_up_proj_weight_quantizers.158 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.gate_up_proj_weight_quantizers.159 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.29e-01 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.gate_up_proj_weight_quantizers.160 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.gate_up_proj_weight_quantizers.161 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.gate_up_proj_weight_quantizers.162 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.26e-01 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.gate_up_proj_weight_quantizers.163 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.gate_up_proj_weight_quantizers.164 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.gate_up_proj_weight_quantizers.165 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.15e-01 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.gate_up_proj_weight_quantizers.166 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.gate_up_proj_weight_quantizers.167 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.gate_up_proj_weight_quantizers.168 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.gate_up_proj_weight_quantizers.169 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.gate_up_proj_weight_quantizers.170 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.gate_up_proj_weight_quantizers.171 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.gate_up_proj_weight_quantizers.172 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.gate_up_proj_weight_quantizers.173 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.gate_up_proj_weight_quantizers.174 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.gate_up_proj_weight_quantizers.175 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.60e-01 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.gate_up_proj_weight_quantizers.176 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.51e-01 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.gate_up_proj_weight_quantizers.177 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.gate_up_proj_weight_quantizers.178 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.gate_up_proj_weight_quantizers.179 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.gate_up_proj_weight_quantizers.180 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.gate_up_proj_weight_quantizers.181 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.16e-01 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.gate_up_proj_weight_quantizers.182 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.gate_up_proj_weight_quantizers.183 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.gate_up_proj_weight_quantizers.184 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.18e-01 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.gate_up_proj_weight_quantizers.185 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.gate_up_proj_weight_quantizers.186 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.97e-01 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.gate_up_proj_weight_quantizers.187 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.63e-01 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.gate_up_proj_weight_quantizers.188 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.86e-01 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.gate_up_proj_weight_quantizers.189 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.gate_up_proj_weight_quantizers.190 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.09e-01 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.gate_up_proj_weight_quantizers.191 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.gate_up_proj_weight_quantizers.192 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.43e-01 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.gate_up_proj_weight_quantizers.193 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.gate_up_proj_weight_quantizers.194 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.gate_up_proj_weight_quantizers.195 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.24e-01 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.gate_up_proj_weight_quantizers.196 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.gate_up_proj_weight_quantizers.197 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.29e-01 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.gate_up_proj_weight_quantizers.198 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.21e-01 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.gate_up_proj_weight_quantizers.199 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.13e-01 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.gate_up_proj_weight_quantizers.200 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.gate_up_proj_weight_quantizers.201 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.gate_up_proj_weight_quantizers.202 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.50e-01 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.gate_up_proj_weight_quantizers.203 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.13e-01 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.gate_up_proj_weight_quantizers.204 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.gate_up_proj_weight_quantizers.205 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.gate_up_proj_weight_quantizers.206 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.gate_up_proj_weight_quantizers.207 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.24e-01 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.gate_up_proj_weight_quantizers.208 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.gate_up_proj_weight_quantizers.209 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.22e-01 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.gate_up_proj_weight_quantizers.210 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.gate_up_proj_weight_quantizers.211 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.gate_up_proj_weight_quantizers.212 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.96e-01 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.gate_up_proj_weight_quantizers.213 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.61e-01 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.gate_up_proj_weight_quantizers.214 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.gate_up_proj_weight_quantizers.215 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.66e-01 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.gate_up_proj_weight_quantizers.216 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.gate_up_proj_weight_quantizers.217 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.gate_up_proj_weight_quantizers.218 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.gate_up_proj_weight_quantizers.219 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.95e-01 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.gate_up_proj_weight_quantizers.220 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.07e-01 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.gate_up_proj_weight_quantizers.221 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.gate_up_proj_weight_quantizers.222 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.gate_up_proj_weight_quantizers.223 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.gate_up_proj_weight_quantizers.224 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.gate_up_proj_weight_quantizers.225 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.gate_up_proj_weight_quantizers.226 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.37e-01 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.gate_up_proj_weight_quantizers.227 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.gate_up_proj_weight_quantizers.228 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.71e-01 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.gate_up_proj_weight_quantizers.229 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.gate_up_proj_weight_quantizers.230 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.30e-01 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.gate_up_proj_weight_quantizers.231 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.gate_up_proj_weight_quantizers.232 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.gate_up_proj_weight_quantizers.233 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.gate_up_proj_weight_quantizers.234 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.gate_up_proj_weight_quantizers.235 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.gate_up_proj_weight_quantizers.236 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.26e-01 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.gate_up_proj_weight_quantizers.237 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.23e-01 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.gate_up_proj_weight_quantizers.238 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.gate_up_proj_weight_quantizers.239 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.gate_up_proj_weight_quantizers.240 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.gate_up_proj_weight_quantizers.241 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.31e-01 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.gate_up_proj_weight_quantizers.242 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.gate_up_proj_weight_quantizers.243 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.54e-01 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.gate_up_proj_weight_quantizers.244 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.28e-01 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.gate_up_proj_weight_quantizers.245 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.gate_up_proj_weight_quantizers.246 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.gate_up_proj_weight_quantizers.247 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.gate_up_proj_weight_quantizers.248 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.gate_up_proj_weight_quantizers.249 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.gate_up_proj_weight_quantizers.250 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.gate_up_proj_weight_quantizers.251 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.19e-01 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.gate_up_proj_weight_quantizers.252 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.63e-01 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.gate_up_proj_weight_quantizers.253 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.gate_up_proj_weight_quantizers.254 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.gate_up_proj_weight_quantizers.255 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.26e-01 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.down_proj_input_quantizer TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.14e+00 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.down_proj_weight_quantizers.0 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.down_proj_weight_quantizers.1 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.down_proj_weight_quantizers.2 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.19e-01 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.down_proj_weight_quantizers.3 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.down_proj_weight_quantizers.4 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.down_proj_weight_quantizers.5 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.28e-01 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.down_proj_weight_quantizers.6 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.down_proj_weight_quantizers.7 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.36e-01 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.down_proj_weight_quantizers.8 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.down_proj_weight_quantizers.9 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.down_proj_weight_quantizers.10 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.down_proj_weight_quantizers.11 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.down_proj_weight_quantizers.12 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.down_proj_weight_quantizers.13 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.down_proj_weight_quantizers.14 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.down_proj_weight_quantizers.15 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.down_proj_weight_quantizers.16 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.down_proj_weight_quantizers.17 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.down_proj_weight_quantizers.18 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.down_proj_weight_quantizers.19 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.down_proj_weight_quantizers.20 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.down_proj_weight_quantizers.21 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.down_proj_weight_quantizers.22 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.15e-01 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.down_proj_weight_quantizers.23 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.94e-01 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.down_proj_weight_quantizers.24 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.62e-01 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.down_proj_weight_quantizers.25 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.down_proj_weight_quantizers.26 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.down_proj_weight_quantizers.27 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.down_proj_weight_quantizers.28 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.down_proj_weight_quantizers.29 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.18e-01 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.down_proj_weight_quantizers.30 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.15e-01 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.down_proj_weight_quantizers.31 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.down_proj_weight_quantizers.32 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.26e-01 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.down_proj_weight_quantizers.33 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.23e-01 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.down_proj_weight_quantizers.34 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.15e-01 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.down_proj_weight_quantizers.35 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.down_proj_weight_quantizers.36 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.85e-01 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.down_proj_weight_quantizers.37 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.18e-01 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.down_proj_weight_quantizers.38 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.down_proj_weight_quantizers.39 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.37e-01 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.down_proj_weight_quantizers.40 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.down_proj_weight_quantizers.41 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.43e-01 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.down_proj_weight_quantizers.42 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.down_proj_weight_quantizers.43 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.down_proj_weight_quantizers.44 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.16e-01 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.down_proj_weight_quantizers.45 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.down_proj_weight_quantizers.46 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.down_proj_weight_quantizers.47 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.down_proj_weight_quantizers.48 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.down_proj_weight_quantizers.49 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.45e-01 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.down_proj_weight_quantizers.50 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.down_proj_weight_quantizers.51 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.32e-01 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.down_proj_weight_quantizers.52 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.down_proj_weight_quantizers.53 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.18e-01 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.down_proj_weight_quantizers.54 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.down_proj_weight_quantizers.55 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.down_proj_weight_quantizers.56 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.15e-01 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.down_proj_weight_quantizers.57 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.down_proj_weight_quantizers.58 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.54e-01 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.down_proj_weight_quantizers.59 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.down_proj_weight_quantizers.60 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.75e-01 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.down_proj_weight_quantizers.61 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.63e-01 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.down_proj_weight_quantizers.62 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.down_proj_weight_quantizers.63 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.70e-01 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.down_proj_weight_quantizers.64 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.34e-01 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.down_proj_weight_quantizers.65 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.38e-01 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.down_proj_weight_quantizers.66 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.down_proj_weight_quantizers.67 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.16e-01 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.down_proj_weight_quantizers.68 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.down_proj_weight_quantizers.69 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.25e-01 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.down_proj_weight_quantizers.70 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.18e-01 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.down_proj_weight_quantizers.71 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.down_proj_weight_quantizers.72 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.down_proj_weight_quantizers.73 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.down_proj_weight_quantizers.74 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.61e-01 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.down_proj_weight_quantizers.75 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.down_proj_weight_quantizers.76 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.down_proj_weight_quantizers.77 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.32e-01 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.down_proj_weight_quantizers.78 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.55e-01 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.down_proj_weight_quantizers.79 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.77e-01 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.down_proj_weight_quantizers.80 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.14e-01 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.down_proj_weight_quantizers.81 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.down_proj_weight_quantizers.82 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.44e-01 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.down_proj_weight_quantizers.83 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.82e-01 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.down_proj_weight_quantizers.84 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.39e-01 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.down_proj_weight_quantizers.85 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.19e-01 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.down_proj_weight_quantizers.86 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.down_proj_weight_quantizers.87 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.down_proj_weight_quantizers.88 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.down_proj_weight_quantizers.89 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.down_proj_weight_quantizers.90 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.36e-01 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.down_proj_weight_quantizers.91 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.down_proj_weight_quantizers.92 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.down_proj_weight_quantizers.93 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.down_proj_weight_quantizers.94 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.24e-01 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.down_proj_weight_quantizers.95 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.30e-01 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.down_proj_weight_quantizers.96 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.down_proj_weight_quantizers.97 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.down_proj_weight_quantizers.98 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.50e-01 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.down_proj_weight_quantizers.99 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.62e-01 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.down_proj_weight_quantizers.100 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.down_proj_weight_quantizers.101 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.32e-01 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.down_proj_weight_quantizers.102 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.down_proj_weight_quantizers.103 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.36e-01 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.down_proj_weight_quantizers.104 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.64e-02 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.down_proj_weight_quantizers.105 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.down_proj_weight_quantizers.106 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.down_proj_weight_quantizers.107 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.down_proj_weight_quantizers.108 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.58e-01 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.down_proj_weight_quantizers.109 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.down_proj_weight_quantizers.110 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.down_proj_weight_quantizers.111 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.down_proj_weight_quantizers.112 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.14e-01 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.down_proj_weight_quantizers.113 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.15e-01 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.down_proj_weight_quantizers.114 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.62e-01 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.down_proj_weight_quantizers.115 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.32e-01 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.down_proj_weight_quantizers.116 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.45e-01 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.down_proj_weight_quantizers.117 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.down_proj_weight_quantizers.118 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.down_proj_weight_quantizers.119 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.22e-01 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.down_proj_weight_quantizers.120 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.down_proj_weight_quantizers.121 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.down_proj_weight_quantizers.122 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.down_proj_weight_quantizers.123 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.down_proj_weight_quantizers.124 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.down_proj_weight_quantizers.125 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.18e-01 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.down_proj_weight_quantizers.126 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.64e-02 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.down_proj_weight_quantizers.127 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.down_proj_weight_quantizers.128 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.24e-01 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.down_proj_weight_quantizers.129 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.down_proj_weight_quantizers.130 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.down_proj_weight_quantizers.131 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.down_proj_weight_quantizers.132 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.down_proj_weight_quantizers.133 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.down_proj_weight_quantizers.134 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.93e-01 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.down_proj_weight_quantizers.135 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.43e-01 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.down_proj_weight_quantizers.136 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.47e-01 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.down_proj_weight_quantizers.137 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.down_proj_weight_quantizers.138 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.down_proj_weight_quantizers.139 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.down_proj_weight_quantizers.140 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.45e-01 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.down_proj_weight_quantizers.141 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.down_proj_weight_quantizers.142 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.down_proj_weight_quantizers.143 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.down_proj_weight_quantizers.144 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.30e-01 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.down_proj_weight_quantizers.145 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.14e-01 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.down_proj_weight_quantizers.146 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.down_proj_weight_quantizers.147 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.down_proj_weight_quantizers.148 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.down_proj_weight_quantizers.149 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.down_proj_weight_quantizers.150 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.down_proj_weight_quantizers.151 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.15e-01 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.down_proj_weight_quantizers.152 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.22e-01 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.down_proj_weight_quantizers.153 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.down_proj_weight_quantizers.154 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.down_proj_weight_quantizers.155 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.down_proj_weight_quantizers.156 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.21e-01 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.down_proj_weight_quantizers.157 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.42e-01 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.down_proj_weight_quantizers.158 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.down_proj_weight_quantizers.159 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.down_proj_weight_quantizers.160 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.down_proj_weight_quantizers.161 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.down_proj_weight_quantizers.162 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.13e-01 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.down_proj_weight_quantizers.163 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.down_proj_weight_quantizers.164 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.down_proj_weight_quantizers.165 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.35e-01 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.down_proj_weight_quantizers.166 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.down_proj_weight_quantizers.167 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.down_proj_weight_quantizers.168 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.36e-01 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.down_proj_weight_quantizers.169 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.down_proj_weight_quantizers.170 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.down_proj_weight_quantizers.171 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.down_proj_weight_quantizers.172 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.down_proj_weight_quantizers.173 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.down_proj_weight_quantizers.174 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.down_proj_weight_quantizers.175 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.18e-01 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.down_proj_weight_quantizers.176 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.44e-01 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.down_proj_weight_quantizers.177 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.down_proj_weight_quantizers.178 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.down_proj_weight_quantizers.179 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.98e-01 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.down_proj_weight_quantizers.180 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.15e-01 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.down_proj_weight_quantizers.181 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.38e-01 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.down_proj_weight_quantizers.182 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.down_proj_weight_quantizers.183 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.54e-01 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.down_proj_weight_quantizers.184 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.30e-01 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.down_proj_weight_quantizers.185 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.down_proj_weight_quantizers.186 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.down_proj_weight_quantizers.187 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.47e-01 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.down_proj_weight_quantizers.188 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.down_proj_weight_quantizers.189 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.down_proj_weight_quantizers.190 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.19e-01 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.down_proj_weight_quantizers.191 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.down_proj_weight_quantizers.192 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.29e-01 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.down_proj_weight_quantizers.193 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.down_proj_weight_quantizers.194 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.down_proj_weight_quantizers.195 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.21e-01 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.down_proj_weight_quantizers.196 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.down_proj_weight_quantizers.197 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.down_proj_weight_quantizers.198 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.down_proj_weight_quantizers.199 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.down_proj_weight_quantizers.200 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.46e-01 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.down_proj_weight_quantizers.201 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.down_proj_weight_quantizers.202 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.68e-01 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.down_proj_weight_quantizers.203 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.23e-01 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.down_proj_weight_quantizers.204 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.down_proj_weight_quantizers.205 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.down_proj_weight_quantizers.206 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.down_proj_weight_quantizers.207 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.down_proj_weight_quantizers.208 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.down_proj_weight_quantizers.209 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.down_proj_weight_quantizers.210 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.down_proj_weight_quantizers.211 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.down_proj_weight_quantizers.212 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.down_proj_weight_quantizers.213 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.down_proj_weight_quantizers.214 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.down_proj_weight_quantizers.215 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.01e-01 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.down_proj_weight_quantizers.216 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.55e-01 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.down_proj_weight_quantizers.217 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.64e-01 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.down_proj_weight_quantizers.218 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.down_proj_weight_quantizers.219 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.down_proj_weight_quantizers.220 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.43e-01 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.down_proj_weight_quantizers.221 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.down_proj_weight_quantizers.222 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.down_proj_weight_quantizers.223 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.down_proj_weight_quantizers.224 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.55e-01 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.down_proj_weight_quantizers.225 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.34e-01 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.down_proj_weight_quantizers.226 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.34e-01 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.down_proj_weight_quantizers.227 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.44e-01 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.down_proj_weight_quantizers.228 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.81e-01 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.down_proj_weight_quantizers.229 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.down_proj_weight_quantizers.230 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.19e-01 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.down_proj_weight_quantizers.231 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.down_proj_weight_quantizers.232 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.down_proj_weight_quantizers.233 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.14e-01 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.down_proj_weight_quantizers.234 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.down_proj_weight_quantizers.235 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.down_proj_weight_quantizers.236 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.27e-01 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.down_proj_weight_quantizers.237 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.25e-01 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.down_proj_weight_quantizers.238 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.down_proj_weight_quantizers.239 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.down_proj_weight_quantizers.240 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.39e-01 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.down_proj_weight_quantizers.241 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.30e-01 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.down_proj_weight_quantizers.242 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.down_proj_weight_quantizers.243 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.26e-01 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.down_proj_weight_quantizers.244 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.95e-01 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.down_proj_weight_quantizers.245 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.down_proj_weight_quantizers.246 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.down_proj_weight_quantizers.247 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.down_proj_weight_quantizers.248 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.down_proj_weight_quantizers.249 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.down_proj_weight_quantizers.250 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.52e-01 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.down_proj_weight_quantizers.251 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.32e-01 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.down_proj_weight_quantizers.252 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.down_proj_weight_quantizers.253 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.down_proj_weight_quantizers.254 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.down_proj_weight_quantizers.255 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.41e-01 calibrator=MaxCalibrator quant) +model.layers.12.mlp.shared_experts.gate_proj.input_quantizer TensorQuantizer(disabled) +model.layers.12.mlp.shared_experts.gate_proj.output_quantizer TensorQuantizer(disabled) +model.layers.12.mlp.shared_experts.gate_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.12.mlp.shared_experts.up_proj.input_quantizer TensorQuantizer(disabled) +model.layers.12.mlp.shared_experts.up_proj.output_quantizer TensorQuantizer(disabled) +model.layers.12.mlp.shared_experts.up_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.12.mlp.shared_experts.down_proj.input_quantizer TensorQuantizer(disabled) +model.layers.12.mlp.shared_experts.down_proj.output_quantizer TensorQuantizer(disabled) +model.layers.12.mlp.shared_experts.down_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.13.self_attn.q_a_proj.input_quantizer TensorQuantizer(disabled) +model.layers.13.self_attn.q_a_proj.output_quantizer TensorQuantizer(disabled) +model.layers.13.self_attn.q_a_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.13.self_attn.q_b_proj.input_quantizer TensorQuantizer(disabled) +model.layers.13.self_attn.q_b_proj.output_quantizer TensorQuantizer(disabled) +model.layers.13.self_attn.q_b_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.13.self_attn.kv_a_proj_with_mqa.input_quantizer TensorQuantizer(disabled) +model.layers.13.self_attn.kv_a_proj_with_mqa.output_quantizer TensorQuantizer(disabled) +model.layers.13.self_attn.kv_a_proj_with_mqa.weight_quantizer TensorQuantizer(disabled) +model.layers.13.self_attn.kv_b_proj.input_quantizer TensorQuantizer(disabled) +model.layers.13.self_attn.kv_b_proj.output_quantizer TensorQuantizer(disabled) +model.layers.13.self_attn.kv_b_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.13.self_attn.o_proj.input_quantizer TensorQuantizer(disabled) +model.layers.13.self_attn.o_proj.output_quantizer TensorQuantizer(disabled) +model.layers.13.self_attn.o_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.13.self_attn.q_bmm_quantizer TensorQuantizer(disabled) +model.layers.13.self_attn.k_bmm_quantizer TensorQuantizer((4, 3) bit fake per-tensor amax=4.48e+02(const) calibrator=MaxCalibrator quant) +model.layers.13.self_attn.v_bmm_quantizer TensorQuantizer((4, 3) bit fake per-tensor amax=4.48e+02(const) calibrator=MaxCalibrator quant) +model.layers.13.self_attn.p_bmm_quantizer TensorQuantizer(disabled) +model.layers.13.mlp.experts.gate_up_proj_input_quantizer TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=5.66e-01 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.gate_up_proj_weight_quantizers.0 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.gate_up_proj_weight_quantizers.1 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.gate_up_proj_weight_quantizers.2 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.gate_up_proj_weight_quantizers.3 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.gate_up_proj_weight_quantizers.4 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.gate_up_proj_weight_quantizers.5 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.gate_up_proj_weight_quantizers.6 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.gate_up_proj_weight_quantizers.7 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.gate_up_proj_weight_quantizers.8 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.gate_up_proj_weight_quantizers.9 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.gate_up_proj_weight_quantizers.10 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.gate_up_proj_weight_quantizers.11 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.gate_up_proj_weight_quantizers.12 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.gate_up_proj_weight_quantizers.13 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.gate_up_proj_weight_quantizers.14 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.gate_up_proj_weight_quantizers.15 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.gate_up_proj_weight_quantizers.16 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.gate_up_proj_weight_quantizers.17 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.gate_up_proj_weight_quantizers.18 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.gate_up_proj_weight_quantizers.19 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.gate_up_proj_weight_quantizers.20 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.gate_up_proj_weight_quantizers.21 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.gate_up_proj_weight_quantizers.22 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.gate_up_proj_weight_quantizers.23 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.gate_up_proj_weight_quantizers.24 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.gate_up_proj_weight_quantizers.25 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.gate_up_proj_weight_quantizers.26 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.gate_up_proj_weight_quantizers.27 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.gate_up_proj_weight_quantizers.28 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.gate_up_proj_weight_quantizers.29 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.gate_up_proj_weight_quantizers.30 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.gate_up_proj_weight_quantizers.31 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.gate_up_proj_weight_quantizers.32 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.gate_up_proj_weight_quantizers.33 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.gate_up_proj_weight_quantizers.34 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.gate_up_proj_weight_quantizers.35 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.gate_up_proj_weight_quantizers.36 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.gate_up_proj_weight_quantizers.37 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.gate_up_proj_weight_quantizers.38 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.gate_up_proj_weight_quantizers.39 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.gate_up_proj_weight_quantizers.40 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.gate_up_proj_weight_quantizers.41 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.gate_up_proj_weight_quantizers.42 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.gate_up_proj_weight_quantizers.43 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.gate_up_proj_weight_quantizers.44 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.gate_up_proj_weight_quantizers.45 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.gate_up_proj_weight_quantizers.46 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.gate_up_proj_weight_quantizers.47 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.gate_up_proj_weight_quantizers.48 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.gate_up_proj_weight_quantizers.49 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.gate_up_proj_weight_quantizers.50 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.gate_up_proj_weight_quantizers.51 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.gate_up_proj_weight_quantizers.52 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.gate_up_proj_weight_quantizers.53 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.gate_up_proj_weight_quantizers.54 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.gate_up_proj_weight_quantizers.55 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.gate_up_proj_weight_quantizers.56 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.gate_up_proj_weight_quantizers.57 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.gate_up_proj_weight_quantizers.58 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.gate_up_proj_weight_quantizers.59 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.gate_up_proj_weight_quantizers.60 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.gate_up_proj_weight_quantizers.61 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.gate_up_proj_weight_quantizers.62 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.gate_up_proj_weight_quantizers.63 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.gate_up_proj_weight_quantizers.64 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.gate_up_proj_weight_quantizers.65 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.gate_up_proj_weight_quantizers.66 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.gate_up_proj_weight_quantizers.67 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.gate_up_proj_weight_quantizers.68 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.gate_up_proj_weight_quantizers.69 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.gate_up_proj_weight_quantizers.70 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.gate_up_proj_weight_quantizers.71 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.gate_up_proj_weight_quantizers.72 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.gate_up_proj_weight_quantizers.73 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.gate_up_proj_weight_quantizers.74 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.gate_up_proj_weight_quantizers.75 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.gate_up_proj_weight_quantizers.76 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.gate_up_proj_weight_quantizers.77 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.18e-01 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.gate_up_proj_weight_quantizers.78 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.gate_up_proj_weight_quantizers.79 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.20e-01 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.gate_up_proj_weight_quantizers.80 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.gate_up_proj_weight_quantizers.81 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.gate_up_proj_weight_quantizers.82 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.gate_up_proj_weight_quantizers.83 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.gate_up_proj_weight_quantizers.84 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.gate_up_proj_weight_quantizers.85 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.gate_up_proj_weight_quantizers.86 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.gate_up_proj_weight_quantizers.87 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.gate_up_proj_weight_quantizers.88 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.gate_up_proj_weight_quantizers.89 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.gate_up_proj_weight_quantizers.90 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.gate_up_proj_weight_quantizers.91 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.90e-01 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.gate_up_proj_weight_quantizers.92 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.gate_up_proj_weight_quantizers.93 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.gate_up_proj_weight_quantizers.94 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.gate_up_proj_weight_quantizers.95 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.gate_up_proj_weight_quantizers.96 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.gate_up_proj_weight_quantizers.97 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.gate_up_proj_weight_quantizers.98 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.gate_up_proj_weight_quantizers.99 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.gate_up_proj_weight_quantizers.100 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.gate_up_proj_weight_quantizers.101 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.gate_up_proj_weight_quantizers.102 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.gate_up_proj_weight_quantizers.103 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.gate_up_proj_weight_quantizers.104 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.gate_up_proj_weight_quantizers.105 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.gate_up_proj_weight_quantizers.106 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.gate_up_proj_weight_quantizers.107 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.gate_up_proj_weight_quantizers.108 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.gate_up_proj_weight_quantizers.109 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.gate_up_proj_weight_quantizers.110 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.gate_up_proj_weight_quantizers.111 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.gate_up_proj_weight_quantizers.112 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.gate_up_proj_weight_quantizers.113 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.gate_up_proj_weight_quantizers.114 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.72e-01 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.gate_up_proj_weight_quantizers.115 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.gate_up_proj_weight_quantizers.116 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.gate_up_proj_weight_quantizers.117 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.gate_up_proj_weight_quantizers.118 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.gate_up_proj_weight_quantizers.119 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.gate_up_proj_weight_quantizers.120 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.gate_up_proj_weight_quantizers.121 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.gate_up_proj_weight_quantizers.122 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.gate_up_proj_weight_quantizers.123 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.gate_up_proj_weight_quantizers.124 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.gate_up_proj_weight_quantizers.125 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.gate_up_proj_weight_quantizers.126 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.gate_up_proj_weight_quantizers.127 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.gate_up_proj_weight_quantizers.128 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.gate_up_proj_weight_quantizers.129 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.gate_up_proj_weight_quantizers.130 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.gate_up_proj_weight_quantizers.131 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.gate_up_proj_weight_quantizers.132 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.gate_up_proj_weight_quantizers.133 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.gate_up_proj_weight_quantizers.134 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.gate_up_proj_weight_quantizers.135 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.gate_up_proj_weight_quantizers.136 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.gate_up_proj_weight_quantizers.137 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.gate_up_proj_weight_quantizers.138 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.gate_up_proj_weight_quantizers.139 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.gate_up_proj_weight_quantizers.140 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.gate_up_proj_weight_quantizers.141 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.gate_up_proj_weight_quantizers.142 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.gate_up_proj_weight_quantizers.143 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.gate_up_proj_weight_quantizers.144 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.gate_up_proj_weight_quantizers.145 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.gate_up_proj_weight_quantizers.146 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.gate_up_proj_weight_quantizers.147 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.gate_up_proj_weight_quantizers.148 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.gate_up_proj_weight_quantizers.149 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.gate_up_proj_weight_quantizers.150 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.gate_up_proj_weight_quantizers.151 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.gate_up_proj_weight_quantizers.152 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.gate_up_proj_weight_quantizers.153 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.gate_up_proj_weight_quantizers.154 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.gate_up_proj_weight_quantizers.155 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.gate_up_proj_weight_quantizers.156 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.gate_up_proj_weight_quantizers.157 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.gate_up_proj_weight_quantizers.158 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.gate_up_proj_weight_quantizers.159 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.gate_up_proj_weight_quantizers.160 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.gate_up_proj_weight_quantizers.161 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.gate_up_proj_weight_quantizers.162 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.gate_up_proj_weight_quantizers.163 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.gate_up_proj_weight_quantizers.164 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.gate_up_proj_weight_quantizers.165 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.gate_up_proj_weight_quantizers.166 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.gate_up_proj_weight_quantizers.167 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.gate_up_proj_weight_quantizers.168 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.gate_up_proj_weight_quantizers.169 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.gate_up_proj_weight_quantizers.170 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.gate_up_proj_weight_quantizers.171 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.gate_up_proj_weight_quantizers.172 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.gate_up_proj_weight_quantizers.173 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.gate_up_proj_weight_quantizers.174 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.gate_up_proj_weight_quantizers.175 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.gate_up_proj_weight_quantizers.176 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.gate_up_proj_weight_quantizers.177 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.gate_up_proj_weight_quantizers.178 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.gate_up_proj_weight_quantizers.179 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.gate_up_proj_weight_quantizers.180 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.gate_up_proj_weight_quantizers.181 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.gate_up_proj_weight_quantizers.182 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.gate_up_proj_weight_quantizers.183 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.gate_up_proj_weight_quantizers.184 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.gate_up_proj_weight_quantizers.185 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.gate_up_proj_weight_quantizers.186 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.gate_up_proj_weight_quantizers.187 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.gate_up_proj_weight_quantizers.188 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.gate_up_proj_weight_quantizers.189 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.gate_up_proj_weight_quantizers.190 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.gate_up_proj_weight_quantizers.191 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.gate_up_proj_weight_quantizers.192 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.gate_up_proj_weight_quantizers.193 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.27e-01 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.gate_up_proj_weight_quantizers.194 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.gate_up_proj_weight_quantizers.195 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.gate_up_proj_weight_quantizers.196 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.gate_up_proj_weight_quantizers.197 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.gate_up_proj_weight_quantizers.198 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.gate_up_proj_weight_quantizers.199 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.gate_up_proj_weight_quantizers.200 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.gate_up_proj_weight_quantizers.201 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.gate_up_proj_weight_quantizers.202 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.09e-01 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.gate_up_proj_weight_quantizers.203 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.gate_up_proj_weight_quantizers.204 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.gate_up_proj_weight_quantizers.205 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.gate_up_proj_weight_quantizers.206 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.gate_up_proj_weight_quantizers.207 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.gate_up_proj_weight_quantizers.208 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.gate_up_proj_weight_quantizers.209 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.gate_up_proj_weight_quantizers.210 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.gate_up_proj_weight_quantizers.211 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.17e-01 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.gate_up_proj_weight_quantizers.212 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.gate_up_proj_weight_quantizers.213 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.gate_up_proj_weight_quantizers.214 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.gate_up_proj_weight_quantizers.215 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.gate_up_proj_weight_quantizers.216 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.gate_up_proj_weight_quantizers.217 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.gate_up_proj_weight_quantizers.218 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.gate_up_proj_weight_quantizers.219 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.gate_up_proj_weight_quantizers.220 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.gate_up_proj_weight_quantizers.221 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.gate_up_proj_weight_quantizers.222 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.gate_up_proj_weight_quantizers.223 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.gate_up_proj_weight_quantizers.224 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.gate_up_proj_weight_quantizers.225 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.gate_up_proj_weight_quantizers.226 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.gate_up_proj_weight_quantizers.227 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.gate_up_proj_weight_quantizers.228 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.gate_up_proj_weight_quantizers.229 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.gate_up_proj_weight_quantizers.230 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.gate_up_proj_weight_quantizers.231 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.gate_up_proj_weight_quantizers.232 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.gate_up_proj_weight_quantizers.233 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.gate_up_proj_weight_quantizers.234 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.gate_up_proj_weight_quantizers.235 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.gate_up_proj_weight_quantizers.236 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.gate_up_proj_weight_quantizers.237 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.gate_up_proj_weight_quantizers.238 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.gate_up_proj_weight_quantizers.239 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.gate_up_proj_weight_quantizers.240 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.gate_up_proj_weight_quantizers.241 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.gate_up_proj_weight_quantizers.242 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.gate_up_proj_weight_quantizers.243 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.gate_up_proj_weight_quantizers.244 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.29e-01 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.gate_up_proj_weight_quantizers.245 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.gate_up_proj_weight_quantizers.246 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.gate_up_proj_weight_quantizers.247 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.gate_up_proj_weight_quantizers.248 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.gate_up_proj_weight_quantizers.249 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.gate_up_proj_weight_quantizers.250 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.gate_up_proj_weight_quantizers.251 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.gate_up_proj_weight_quantizers.252 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.gate_up_proj_weight_quantizers.253 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.gate_up_proj_weight_quantizers.254 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.gate_up_proj_weight_quantizers.255 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.down_proj_input_quantizer TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.72e+00 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.down_proj_weight_quantizers.0 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.down_proj_weight_quantizers.1 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.down_proj_weight_quantizers.2 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.23e-01 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.down_proj_weight_quantizers.3 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.down_proj_weight_quantizers.4 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.down_proj_weight_quantizers.5 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.down_proj_weight_quantizers.6 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.down_proj_weight_quantizers.7 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.down_proj_weight_quantizers.8 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.down_proj_weight_quantizers.9 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.down_proj_weight_quantizers.10 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.down_proj_weight_quantizers.11 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.down_proj_weight_quantizers.12 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.down_proj_weight_quantizers.13 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.down_proj_weight_quantizers.14 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.down_proj_weight_quantizers.15 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.down_proj_weight_quantizers.16 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.down_proj_weight_quantizers.17 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.down_proj_weight_quantizers.18 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.down_proj_weight_quantizers.19 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.13e-01 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.down_proj_weight_quantizers.20 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.down_proj_weight_quantizers.21 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.28e-01 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.down_proj_weight_quantizers.22 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.down_proj_weight_quantizers.23 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.down_proj_weight_quantizers.24 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.down_proj_weight_quantizers.25 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.down_proj_weight_quantizers.26 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.down_proj_weight_quantizers.27 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.down_proj_weight_quantizers.28 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.down_proj_weight_quantizers.29 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.down_proj_weight_quantizers.30 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.down_proj_weight_quantizers.31 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.down_proj_weight_quantizers.32 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.down_proj_weight_quantizers.33 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.down_proj_weight_quantizers.34 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.down_proj_weight_quantizers.35 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.down_proj_weight_quantizers.36 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.down_proj_weight_quantizers.37 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.down_proj_weight_quantizers.38 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.down_proj_weight_quantizers.39 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.down_proj_weight_quantizers.40 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.down_proj_weight_quantizers.41 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.down_proj_weight_quantizers.42 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.down_proj_weight_quantizers.43 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.down_proj_weight_quantizers.44 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.33e-01 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.down_proj_weight_quantizers.45 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.down_proj_weight_quantizers.46 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.down_proj_weight_quantizers.47 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.down_proj_weight_quantizers.48 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.down_proj_weight_quantizers.49 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.down_proj_weight_quantizers.50 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.34e-01 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.down_proj_weight_quantizers.51 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.down_proj_weight_quantizers.52 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.down_proj_weight_quantizers.53 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.down_proj_weight_quantizers.54 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.down_proj_weight_quantizers.55 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.64e-01 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.down_proj_weight_quantizers.56 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.40e-01 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.down_proj_weight_quantizers.57 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.down_proj_weight_quantizers.58 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.down_proj_weight_quantizers.59 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.down_proj_weight_quantizers.60 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.21e-01 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.down_proj_weight_quantizers.61 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.down_proj_weight_quantizers.62 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.down_proj_weight_quantizers.63 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.down_proj_weight_quantizers.64 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.14e-01 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.down_proj_weight_quantizers.65 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.36e-01 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.down_proj_weight_quantizers.66 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.24e-01 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.down_proj_weight_quantizers.67 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.down_proj_weight_quantizers.68 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.down_proj_weight_quantizers.69 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.down_proj_weight_quantizers.70 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.down_proj_weight_quantizers.71 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.down_proj_weight_quantizers.72 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.down_proj_weight_quantizers.73 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.down_proj_weight_quantizers.74 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.14e-01 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.down_proj_weight_quantizers.75 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.down_proj_weight_quantizers.76 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.down_proj_weight_quantizers.77 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.38e-01 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.down_proj_weight_quantizers.78 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.down_proj_weight_quantizers.79 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.down_proj_weight_quantizers.80 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.down_proj_weight_quantizers.81 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.down_proj_weight_quantizers.82 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.down_proj_weight_quantizers.83 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.down_proj_weight_quantizers.84 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.down_proj_weight_quantizers.85 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.down_proj_weight_quantizers.86 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.13e-01 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.down_proj_weight_quantizers.87 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.27e-01 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.down_proj_weight_quantizers.88 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.down_proj_weight_quantizers.89 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.down_proj_weight_quantizers.90 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.down_proj_weight_quantizers.91 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.98e-01 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.down_proj_weight_quantizers.92 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.down_proj_weight_quantizers.93 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.down_proj_weight_quantizers.94 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.down_proj_weight_quantizers.95 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.down_proj_weight_quantizers.96 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.down_proj_weight_quantizers.97 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.down_proj_weight_quantizers.98 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.47e-01 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.down_proj_weight_quantizers.99 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.down_proj_weight_quantizers.100 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.down_proj_weight_quantizers.101 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.down_proj_weight_quantizers.102 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.30e-01 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.down_proj_weight_quantizers.103 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.down_proj_weight_quantizers.104 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.down_proj_weight_quantizers.105 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.down_proj_weight_quantizers.106 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.down_proj_weight_quantizers.107 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.down_proj_weight_quantizers.108 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.down_proj_weight_quantizers.109 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.down_proj_weight_quantizers.110 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.down_proj_weight_quantizers.111 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.down_proj_weight_quantizers.112 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.down_proj_weight_quantizers.113 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.down_proj_weight_quantizers.114 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.16e-01 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.down_proj_weight_quantizers.115 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.down_proj_weight_quantizers.116 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.down_proj_weight_quantizers.117 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.down_proj_weight_quantizers.118 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.down_proj_weight_quantizers.119 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.down_proj_weight_quantizers.120 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.down_proj_weight_quantizers.121 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.15e-01 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.down_proj_weight_quantizers.122 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.down_proj_weight_quantizers.123 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.down_proj_weight_quantizers.124 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.down_proj_weight_quantizers.125 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.down_proj_weight_quantizers.126 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.down_proj_weight_quantizers.127 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.30e-01 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.down_proj_weight_quantizers.128 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.down_proj_weight_quantizers.129 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.down_proj_weight_quantizers.130 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.down_proj_weight_quantizers.131 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.26e-01 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.down_proj_weight_quantizers.132 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.down_proj_weight_quantizers.133 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.down_proj_weight_quantizers.134 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.down_proj_weight_quantizers.135 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.14e-01 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.down_proj_weight_quantizers.136 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.down_proj_weight_quantizers.137 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.down_proj_weight_quantizers.138 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.down_proj_weight_quantizers.139 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.36e-01 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.down_proj_weight_quantizers.140 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.down_proj_weight_quantizers.141 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.down_proj_weight_quantizers.142 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.down_proj_weight_quantizers.143 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.down_proj_weight_quantizers.144 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.down_proj_weight_quantizers.145 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.16e-01 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.down_proj_weight_quantizers.146 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.18e-01 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.down_proj_weight_quantizers.147 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.down_proj_weight_quantizers.148 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.down_proj_weight_quantizers.149 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.down_proj_weight_quantizers.150 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.down_proj_weight_quantizers.151 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.down_proj_weight_quantizers.152 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.down_proj_weight_quantizers.153 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.down_proj_weight_quantizers.154 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.down_proj_weight_quantizers.155 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.down_proj_weight_quantizers.156 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.down_proj_weight_quantizers.157 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.down_proj_weight_quantizers.158 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.23e-01 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.down_proj_weight_quantizers.159 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.down_proj_weight_quantizers.160 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.down_proj_weight_quantizers.161 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.down_proj_weight_quantizers.162 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.down_proj_weight_quantizers.163 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.down_proj_weight_quantizers.164 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.20e-01 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.down_proj_weight_quantizers.165 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.down_proj_weight_quantizers.166 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.19e-01 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.down_proj_weight_quantizers.167 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.down_proj_weight_quantizers.168 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.down_proj_weight_quantizers.169 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.down_proj_weight_quantizers.170 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.22e-01 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.down_proj_weight_quantizers.171 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.down_proj_weight_quantizers.172 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.26e-01 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.down_proj_weight_quantizers.173 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.down_proj_weight_quantizers.174 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.down_proj_weight_quantizers.175 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.down_proj_weight_quantizers.176 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.down_proj_weight_quantizers.177 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.down_proj_weight_quantizers.178 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.down_proj_weight_quantizers.179 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.29e-01 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.down_proj_weight_quantizers.180 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.down_proj_weight_quantizers.181 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.down_proj_weight_quantizers.182 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.down_proj_weight_quantizers.183 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.14e-01 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.down_proj_weight_quantizers.184 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.down_proj_weight_quantizers.185 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.14e-01 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.down_proj_weight_quantizers.186 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.39e-01 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.down_proj_weight_quantizers.187 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.down_proj_weight_quantizers.188 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.down_proj_weight_quantizers.189 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.down_proj_weight_quantizers.190 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.73e-01 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.down_proj_weight_quantizers.191 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.down_proj_weight_quantizers.192 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.down_proj_weight_quantizers.193 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.63e-01 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.down_proj_weight_quantizers.194 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.down_proj_weight_quantizers.195 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.19e-01 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.down_proj_weight_quantizers.196 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.14e-01 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.down_proj_weight_quantizers.197 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.93e-01 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.down_proj_weight_quantizers.198 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.down_proj_weight_quantizers.199 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.down_proj_weight_quantizers.200 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.45e-01 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.down_proj_weight_quantizers.201 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.down_proj_weight_quantizers.202 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.56e-01 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.down_proj_weight_quantizers.203 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.30e-01 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.down_proj_weight_quantizers.204 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.down_proj_weight_quantizers.205 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.down_proj_weight_quantizers.206 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.down_proj_weight_quantizers.207 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.down_proj_weight_quantizers.208 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.33e-01 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.down_proj_weight_quantizers.209 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.down_proj_weight_quantizers.210 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.down_proj_weight_quantizers.211 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.down_proj_weight_quantizers.212 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.down_proj_weight_quantizers.213 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.down_proj_weight_quantizers.214 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.down_proj_weight_quantizers.215 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.down_proj_weight_quantizers.216 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.down_proj_weight_quantizers.217 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.down_proj_weight_quantizers.218 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.down_proj_weight_quantizers.219 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.down_proj_weight_quantizers.220 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.18e-01 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.down_proj_weight_quantizers.221 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.down_proj_weight_quantizers.222 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.down_proj_weight_quantizers.223 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.down_proj_weight_quantizers.224 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.41e-01 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.down_proj_weight_quantizers.225 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.59e-02 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.down_proj_weight_quantizers.226 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.down_proj_weight_quantizers.227 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.down_proj_weight_quantizers.228 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.down_proj_weight_quantizers.229 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.down_proj_weight_quantizers.230 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.down_proj_weight_quantizers.231 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.41e-01 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.down_proj_weight_quantizers.232 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.down_proj_weight_quantizers.233 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.down_proj_weight_quantizers.234 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.54e-01 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.down_proj_weight_quantizers.235 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.down_proj_weight_quantizers.236 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.down_proj_weight_quantizers.237 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.down_proj_weight_quantizers.238 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.down_proj_weight_quantizers.239 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.16e-01 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.down_proj_weight_quantizers.240 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.47e-01 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.down_proj_weight_quantizers.241 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.down_proj_weight_quantizers.242 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.21e-01 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.down_proj_weight_quantizers.243 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.down_proj_weight_quantizers.244 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.down_proj_weight_quantizers.245 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.down_proj_weight_quantizers.246 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.down_proj_weight_quantizers.247 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.down_proj_weight_quantizers.248 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.down_proj_weight_quantizers.249 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.37e-01 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.down_proj_weight_quantizers.250 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.down_proj_weight_quantizers.251 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.down_proj_weight_quantizers.252 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.down_proj_weight_quantizers.253 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.25e-01 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.down_proj_weight_quantizers.254 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.down_proj_weight_quantizers.255 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.13.mlp.shared_experts.gate_proj.input_quantizer TensorQuantizer(disabled) +model.layers.13.mlp.shared_experts.gate_proj.output_quantizer TensorQuantizer(disabled) +model.layers.13.mlp.shared_experts.gate_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.13.mlp.shared_experts.up_proj.input_quantizer TensorQuantizer(disabled) +model.layers.13.mlp.shared_experts.up_proj.output_quantizer TensorQuantizer(disabled) +model.layers.13.mlp.shared_experts.up_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.13.mlp.shared_experts.down_proj.input_quantizer TensorQuantizer(disabled) +model.layers.13.mlp.shared_experts.down_proj.output_quantizer TensorQuantizer(disabled) +model.layers.13.mlp.shared_experts.down_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.14.self_attn.q_a_proj.input_quantizer TensorQuantizer(disabled) +model.layers.14.self_attn.q_a_proj.output_quantizer TensorQuantizer(disabled) +model.layers.14.self_attn.q_a_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.14.self_attn.q_b_proj.input_quantizer TensorQuantizer(disabled) +model.layers.14.self_attn.q_b_proj.output_quantizer TensorQuantizer(disabled) +model.layers.14.self_attn.q_b_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.14.self_attn.kv_a_proj_with_mqa.input_quantizer TensorQuantizer(disabled) +model.layers.14.self_attn.kv_a_proj_with_mqa.output_quantizer TensorQuantizer(disabled) +model.layers.14.self_attn.kv_a_proj_with_mqa.weight_quantizer TensorQuantizer(disabled) +model.layers.14.self_attn.kv_b_proj.input_quantizer TensorQuantizer(disabled) +model.layers.14.self_attn.kv_b_proj.output_quantizer TensorQuantizer(disabled) +model.layers.14.self_attn.kv_b_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.14.self_attn.o_proj.input_quantizer TensorQuantizer(disabled) +model.layers.14.self_attn.o_proj.output_quantizer TensorQuantizer(disabled) +model.layers.14.self_attn.o_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.14.self_attn.indexer.wq_b.input_quantizer TensorQuantizer(disabled) +model.layers.14.self_attn.indexer.wq_b.output_quantizer TensorQuantizer(disabled) +model.layers.14.self_attn.indexer.wq_b.weight_quantizer TensorQuantizer(disabled) +model.layers.14.self_attn.indexer.wk.input_quantizer TensorQuantizer(disabled) +model.layers.14.self_attn.indexer.wk.output_quantizer TensorQuantizer(disabled) +model.layers.14.self_attn.indexer.wk.weight_quantizer TensorQuantizer(disabled) +model.layers.14.self_attn.indexer.k_norm.input_quantizer TensorQuantizer(disabled) +model.layers.14.self_attn.indexer.k_norm.output_quantizer TensorQuantizer(disabled) +model.layers.14.self_attn.indexer.weights_proj.input_quantizer TensorQuantizer(disabled) +model.layers.14.self_attn.indexer.weights_proj.output_quantizer TensorQuantizer(disabled) +model.layers.14.self_attn.indexer.weights_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.14.self_attn.q_bmm_quantizer TensorQuantizer(disabled) +model.layers.14.self_attn.k_bmm_quantizer TensorQuantizer((4, 3) bit fake per-tensor amax=4.48e+02(const) calibrator=MaxCalibrator quant) +model.layers.14.self_attn.v_bmm_quantizer TensorQuantizer((4, 3) bit fake per-tensor amax=4.48e+02(const) calibrator=MaxCalibrator quant) +model.layers.14.self_attn.p_bmm_quantizer TensorQuantizer(disabled) +model.layers.14.mlp.experts.gate_up_proj_input_quantizer TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=5.43e-01 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.gate_up_proj_weight_quantizers.0 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.gate_up_proj_weight_quantizers.1 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.gate_up_proj_weight_quantizers.2 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.gate_up_proj_weight_quantizers.3 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.gate_up_proj_weight_quantizers.4 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.gate_up_proj_weight_quantizers.5 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.gate_up_proj_weight_quantizers.6 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.gate_up_proj_weight_quantizers.7 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.gate_up_proj_weight_quantizers.8 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.gate_up_proj_weight_quantizers.9 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.gate_up_proj_weight_quantizers.10 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.gate_up_proj_weight_quantizers.11 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.gate_up_proj_weight_quantizers.12 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.gate_up_proj_weight_quantizers.13 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.gate_up_proj_weight_quantizers.14 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.gate_up_proj_weight_quantizers.15 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.gate_up_proj_weight_quantizers.16 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.gate_up_proj_weight_quantizers.17 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.gate_up_proj_weight_quantizers.18 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.gate_up_proj_weight_quantizers.19 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=5.98e-01 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.gate_up_proj_weight_quantizers.20 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.gate_up_proj_weight_quantizers.21 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.gate_up_proj_weight_quantizers.22 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.gate_up_proj_weight_quantizers.23 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.gate_up_proj_weight_quantizers.24 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.gate_up_proj_weight_quantizers.25 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.gate_up_proj_weight_quantizers.26 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.gate_up_proj_weight_quantizers.27 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.gate_up_proj_weight_quantizers.28 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=4.63e-01 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.gate_up_proj_weight_quantizers.29 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.gate_up_proj_weight_quantizers.30 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.gate_up_proj_weight_quantizers.31 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.gate_up_proj_weight_quantizers.32 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.gate_up_proj_weight_quantizers.33 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.gate_up_proj_weight_quantizers.34 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.gate_up_proj_weight_quantizers.35 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.gate_up_proj_weight_quantizers.36 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.gate_up_proj_weight_quantizers.37 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.gate_up_proj_weight_quantizers.38 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.gate_up_proj_weight_quantizers.39 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.gate_up_proj_weight_quantizers.40 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.gate_up_proj_weight_quantizers.41 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.gate_up_proj_weight_quantizers.42 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.gate_up_proj_weight_quantizers.43 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.gate_up_proj_weight_quantizers.44 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.gate_up_proj_weight_quantizers.45 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.gate_up_proj_weight_quantizers.46 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.gate_up_proj_weight_quantizers.47 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.43e-01 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.gate_up_proj_weight_quantizers.48 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.gate_up_proj_weight_quantizers.49 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.gate_up_proj_weight_quantizers.50 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.gate_up_proj_weight_quantizers.51 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.31e-01 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.gate_up_proj_weight_quantizers.52 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.gate_up_proj_weight_quantizers.53 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.gate_up_proj_weight_quantizers.54 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.gate_up_proj_weight_quantizers.55 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.35e-01 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.gate_up_proj_weight_quantizers.56 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.gate_up_proj_weight_quantizers.57 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.gate_up_proj_weight_quantizers.58 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.gate_up_proj_weight_quantizers.59 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.gate_up_proj_weight_quantizers.60 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.gate_up_proj_weight_quantizers.61 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.gate_up_proj_weight_quantizers.62 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.gate_up_proj_weight_quantizers.63 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.gate_up_proj_weight_quantizers.64 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.gate_up_proj_weight_quantizers.65 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.gate_up_proj_weight_quantizers.66 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.gate_up_proj_weight_quantizers.67 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.gate_up_proj_weight_quantizers.68 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.gate_up_proj_weight_quantizers.69 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.42e-01 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.gate_up_proj_weight_quantizers.70 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.gate_up_proj_weight_quantizers.71 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.gate_up_proj_weight_quantizers.72 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.gate_up_proj_weight_quantizers.73 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.gate_up_proj_weight_quantizers.74 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.gate_up_proj_weight_quantizers.75 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.gate_up_proj_weight_quantizers.76 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.gate_up_proj_weight_quantizers.77 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.64e-01 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.gate_up_proj_weight_quantizers.78 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.gate_up_proj_weight_quantizers.79 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.gate_up_proj_weight_quantizers.80 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.gate_up_proj_weight_quantizers.81 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.gate_up_proj_weight_quantizers.82 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.gate_up_proj_weight_quantizers.83 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.gate_up_proj_weight_quantizers.84 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.gate_up_proj_weight_quantizers.85 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.gate_up_proj_weight_quantizers.86 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.34e-01 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.gate_up_proj_weight_quantizers.87 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.gate_up_proj_weight_quantizers.88 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.15e-01 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.gate_up_proj_weight_quantizers.89 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.gate_up_proj_weight_quantizers.90 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.gate_up_proj_weight_quantizers.91 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.gate_up_proj_weight_quantizers.92 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.gate_up_proj_weight_quantizers.93 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.03e-01 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.gate_up_proj_weight_quantizers.94 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.gate_up_proj_weight_quantizers.95 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.gate_up_proj_weight_quantizers.96 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.gate_up_proj_weight_quantizers.97 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.gate_up_proj_weight_quantizers.98 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.gate_up_proj_weight_quantizers.99 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.gate_up_proj_weight_quantizers.100 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.gate_up_proj_weight_quantizers.101 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.22e-01 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.gate_up_proj_weight_quantizers.102 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.gate_up_proj_weight_quantizers.103 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.gate_up_proj_weight_quantizers.104 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.gate_up_proj_weight_quantizers.105 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.gate_up_proj_weight_quantizers.106 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.gate_up_proj_weight_quantizers.107 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.gate_up_proj_weight_quantizers.108 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.gate_up_proj_weight_quantizers.109 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.gate_up_proj_weight_quantizers.110 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.gate_up_proj_weight_quantizers.111 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.gate_up_proj_weight_quantizers.112 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.gate_up_proj_weight_quantizers.113 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.gate_up_proj_weight_quantizers.114 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.gate_up_proj_weight_quantizers.115 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.gate_up_proj_weight_quantizers.116 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.60e-01 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.gate_up_proj_weight_quantizers.117 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.gate_up_proj_weight_quantizers.118 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.gate_up_proj_weight_quantizers.119 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.gate_up_proj_weight_quantizers.120 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.gate_up_proj_weight_quantizers.121 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.gate_up_proj_weight_quantizers.122 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.gate_up_proj_weight_quantizers.123 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.gate_up_proj_weight_quantizers.124 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.91e-01 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.gate_up_proj_weight_quantizers.125 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.17e-01 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.gate_up_proj_weight_quantizers.126 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.gate_up_proj_weight_quantizers.127 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.gate_up_proj_weight_quantizers.128 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.gate_up_proj_weight_quantizers.129 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.gate_up_proj_weight_quantizers.130 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.gate_up_proj_weight_quantizers.131 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.gate_up_proj_weight_quantizers.132 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.gate_up_proj_weight_quantizers.133 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.gate_up_proj_weight_quantizers.134 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.gate_up_proj_weight_quantizers.135 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.gate_up_proj_weight_quantizers.136 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.gate_up_proj_weight_quantizers.137 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.gate_up_proj_weight_quantizers.138 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.gate_up_proj_weight_quantizers.139 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.gate_up_proj_weight_quantizers.140 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.gate_up_proj_weight_quantizers.141 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.gate_up_proj_weight_quantizers.142 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.gate_up_proj_weight_quantizers.143 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.gate_up_proj_weight_quantizers.144 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.40e-01 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.gate_up_proj_weight_quantizers.145 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.gate_up_proj_weight_quantizers.146 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.gate_up_proj_weight_quantizers.147 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.gate_up_proj_weight_quantizers.148 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.24e-01 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.gate_up_proj_weight_quantizers.149 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.gate_up_proj_weight_quantizers.150 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.gate_up_proj_weight_quantizers.151 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.gate_up_proj_weight_quantizers.152 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.gate_up_proj_weight_quantizers.153 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.gate_up_proj_weight_quantizers.154 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.02e-01 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.gate_up_proj_weight_quantizers.155 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.17e-01 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.gate_up_proj_weight_quantizers.156 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.gate_up_proj_weight_quantizers.157 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.gate_up_proj_weight_quantizers.158 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.gate_up_proj_weight_quantizers.159 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.gate_up_proj_weight_quantizers.160 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.52e-01 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.gate_up_proj_weight_quantizers.161 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.gate_up_proj_weight_quantizers.162 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.gate_up_proj_weight_quantizers.163 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.gate_up_proj_weight_quantizers.164 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.gate_up_proj_weight_quantizers.165 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.gate_up_proj_weight_quantizers.166 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.gate_up_proj_weight_quantizers.167 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.gate_up_proj_weight_quantizers.168 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.gate_up_proj_weight_quantizers.169 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.gate_up_proj_weight_quantizers.170 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.gate_up_proj_weight_quantizers.171 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.gate_up_proj_weight_quantizers.172 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.gate_up_proj_weight_quantizers.173 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.gate_up_proj_weight_quantizers.174 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.gate_up_proj_weight_quantizers.175 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.gate_up_proj_weight_quantizers.176 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.gate_up_proj_weight_quantizers.177 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.gate_up_proj_weight_quantizers.178 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.gate_up_proj_weight_quantizers.179 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.gate_up_proj_weight_quantizers.180 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.gate_up_proj_weight_quantizers.181 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.gate_up_proj_weight_quantizers.182 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.gate_up_proj_weight_quantizers.183 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.gate_up_proj_weight_quantizers.184 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.gate_up_proj_weight_quantizers.185 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.gate_up_proj_weight_quantizers.186 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.gate_up_proj_weight_quantizers.187 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.13e-01 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.gate_up_proj_weight_quantizers.188 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.79e-01 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.gate_up_proj_weight_quantizers.189 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.gate_up_proj_weight_quantizers.190 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.gate_up_proj_weight_quantizers.191 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.56e-01 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.gate_up_proj_weight_quantizers.192 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.gate_up_proj_weight_quantizers.193 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.gate_up_proj_weight_quantizers.194 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.gate_up_proj_weight_quantizers.195 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.gate_up_proj_weight_quantizers.196 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.gate_up_proj_weight_quantizers.197 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.gate_up_proj_weight_quantizers.198 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.gate_up_proj_weight_quantizers.199 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.gate_up_proj_weight_quantizers.200 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.gate_up_proj_weight_quantizers.201 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.gate_up_proj_weight_quantizers.202 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.gate_up_proj_weight_quantizers.203 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.gate_up_proj_weight_quantizers.204 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.36e-01 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.gate_up_proj_weight_quantizers.205 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.gate_up_proj_weight_quantizers.206 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.gate_up_proj_weight_quantizers.207 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.gate_up_proj_weight_quantizers.208 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.gate_up_proj_weight_quantizers.209 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.gate_up_proj_weight_quantizers.210 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.gate_up_proj_weight_quantizers.211 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.gate_up_proj_weight_quantizers.212 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.gate_up_proj_weight_quantizers.213 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.78e-01 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.gate_up_proj_weight_quantizers.214 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.gate_up_proj_weight_quantizers.215 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.gate_up_proj_weight_quantizers.216 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.53e-01 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.gate_up_proj_weight_quantizers.217 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.gate_up_proj_weight_quantizers.218 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.gate_up_proj_weight_quantizers.219 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.24e-01 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.gate_up_proj_weight_quantizers.220 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.gate_up_proj_weight_quantizers.221 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.19e-01 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.gate_up_proj_weight_quantizers.222 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.gate_up_proj_weight_quantizers.223 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.gate_up_proj_weight_quantizers.224 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.gate_up_proj_weight_quantizers.225 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.gate_up_proj_weight_quantizers.226 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.gate_up_proj_weight_quantizers.227 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.38e-01 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.gate_up_proj_weight_quantizers.228 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.gate_up_proj_weight_quantizers.229 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.gate_up_proj_weight_quantizers.230 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.gate_up_proj_weight_quantizers.231 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.gate_up_proj_weight_quantizers.232 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.gate_up_proj_weight_quantizers.233 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.gate_up_proj_weight_quantizers.234 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.23e-01 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.gate_up_proj_weight_quantizers.235 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.gate_up_proj_weight_quantizers.236 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.gate_up_proj_weight_quantizers.237 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.gate_up_proj_weight_quantizers.238 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.gate_up_proj_weight_quantizers.239 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.gate_up_proj_weight_quantizers.240 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.gate_up_proj_weight_quantizers.241 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.gate_up_proj_weight_quantizers.242 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.46e-01 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.gate_up_proj_weight_quantizers.243 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.gate_up_proj_weight_quantizers.244 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.gate_up_proj_weight_quantizers.245 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.gate_up_proj_weight_quantizers.246 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.56e-01 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.gate_up_proj_weight_quantizers.247 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.gate_up_proj_weight_quantizers.248 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.gate_up_proj_weight_quantizers.249 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.gate_up_proj_weight_quantizers.250 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.gate_up_proj_weight_quantizers.251 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.gate_up_proj_weight_quantizers.252 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.gate_up_proj_weight_quantizers.253 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.gate_up_proj_weight_quantizers.254 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.gate_up_proj_weight_quantizers.255 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.down_proj_input_quantizer TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=4.38e+00 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.down_proj_weight_quantizers.0 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.down_proj_weight_quantizers.1 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.down_proj_weight_quantizers.2 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.down_proj_weight_quantizers.3 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.down_proj_weight_quantizers.4 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.72e-01 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.down_proj_weight_quantizers.5 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.down_proj_weight_quantizers.6 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.down_proj_weight_quantizers.7 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.down_proj_weight_quantizers.8 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.21e-01 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.down_proj_weight_quantizers.9 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.down_proj_weight_quantizers.10 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.down_proj_weight_quantizers.11 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.down_proj_weight_quantizers.12 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.down_proj_weight_quantizers.13 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.down_proj_weight_quantizers.14 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.down_proj_weight_quantizers.15 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.down_proj_weight_quantizers.16 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.down_proj_weight_quantizers.17 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.down_proj_weight_quantizers.18 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.down_proj_weight_quantizers.19 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.14e-01 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.down_proj_weight_quantizers.20 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.down_proj_weight_quantizers.21 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.91e-01 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.down_proj_weight_quantizers.22 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.down_proj_weight_quantizers.23 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.down_proj_weight_quantizers.24 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.down_proj_weight_quantizers.25 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.down_proj_weight_quantizers.26 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.13e-01 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.down_proj_weight_quantizers.27 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.down_proj_weight_quantizers.28 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.50e-01 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.down_proj_weight_quantizers.29 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.down_proj_weight_quantizers.30 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.down_proj_weight_quantizers.31 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.down_proj_weight_quantizers.32 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.14e-01 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.down_proj_weight_quantizers.33 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.59e-02 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.down_proj_weight_quantizers.34 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.64e-02 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.down_proj_weight_quantizers.35 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.down_proj_weight_quantizers.36 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.down_proj_weight_quantizers.37 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.down_proj_weight_quantizers.38 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.15e-01 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.down_proj_weight_quantizers.39 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.45e-02 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.down_proj_weight_quantizers.40 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.down_proj_weight_quantizers.41 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.down_proj_weight_quantizers.42 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.down_proj_weight_quantizers.43 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.down_proj_weight_quantizers.44 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.down_proj_weight_quantizers.45 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.17e-01 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.down_proj_weight_quantizers.46 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.16e-01 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.down_proj_weight_quantizers.47 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.55e-01 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.down_proj_weight_quantizers.48 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.down_proj_weight_quantizers.49 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.down_proj_weight_quantizers.50 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.down_proj_weight_quantizers.51 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.down_proj_weight_quantizers.52 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.down_proj_weight_quantizers.53 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.down_proj_weight_quantizers.54 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.down_proj_weight_quantizers.55 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.down_proj_weight_quantizers.56 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.down_proj_weight_quantizers.57 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.down_proj_weight_quantizers.58 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.37e-01 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.down_proj_weight_quantizers.59 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.down_proj_weight_quantizers.60 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.down_proj_weight_quantizers.61 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.down_proj_weight_quantizers.62 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.down_proj_weight_quantizers.63 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.35e-01 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.down_proj_weight_quantizers.64 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.down_proj_weight_quantizers.65 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.down_proj_weight_quantizers.66 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.down_proj_weight_quantizers.67 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.down_proj_weight_quantizers.68 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.down_proj_weight_quantizers.69 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.80e-01 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.down_proj_weight_quantizers.70 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.72e-01 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.down_proj_weight_quantizers.71 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.down_proj_weight_quantizers.72 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.down_proj_weight_quantizers.73 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.down_proj_weight_quantizers.74 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.down_proj_weight_quantizers.75 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.down_proj_weight_quantizers.76 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.down_proj_weight_quantizers.77 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.down_proj_weight_quantizers.78 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.14e-01 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.down_proj_weight_quantizers.79 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.down_proj_weight_quantizers.80 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.down_proj_weight_quantizers.81 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.down_proj_weight_quantizers.82 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.49e-01 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.down_proj_weight_quantizers.83 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.down_proj_weight_quantizers.84 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.down_proj_weight_quantizers.85 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.down_proj_weight_quantizers.86 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.down_proj_weight_quantizers.87 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.down_proj_weight_quantizers.88 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.down_proj_weight_quantizers.89 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.down_proj_weight_quantizers.90 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.down_proj_weight_quantizers.91 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.down_proj_weight_quantizers.92 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.down_proj_weight_quantizers.93 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.23e-01 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.down_proj_weight_quantizers.94 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.down_proj_weight_quantizers.95 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.down_proj_weight_quantizers.96 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.down_proj_weight_quantizers.97 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.down_proj_weight_quantizers.98 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.down_proj_weight_quantizers.99 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.down_proj_weight_quantizers.100 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.down_proj_weight_quantizers.101 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.94e-01 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.down_proj_weight_quantizers.102 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.down_proj_weight_quantizers.103 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.54e-02 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.down_proj_weight_quantizers.104 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.down_proj_weight_quantizers.105 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.down_proj_weight_quantizers.106 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.down_proj_weight_quantizers.107 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.39e-01 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.down_proj_weight_quantizers.108 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.down_proj_weight_quantizers.109 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.down_proj_weight_quantizers.110 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.down_proj_weight_quantizers.111 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.down_proj_weight_quantizers.112 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.down_proj_weight_quantizers.113 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.18e-01 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.down_proj_weight_quantizers.114 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.down_proj_weight_quantizers.115 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.91e-01 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.down_proj_weight_quantizers.116 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.17e-01 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.down_proj_weight_quantizers.117 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.down_proj_weight_quantizers.118 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.down_proj_weight_quantizers.119 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.58e-01 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.down_proj_weight_quantizers.120 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.down_proj_weight_quantizers.121 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.73e-01 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.down_proj_weight_quantizers.122 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.down_proj_weight_quantizers.123 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.down_proj_weight_quantizers.124 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.07e-01 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.down_proj_weight_quantizers.125 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.59e-01 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.down_proj_weight_quantizers.126 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.down_proj_weight_quantizers.127 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.down_proj_weight_quantizers.128 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.down_proj_weight_quantizers.129 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.down_proj_weight_quantizers.130 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.down_proj_weight_quantizers.131 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.down_proj_weight_quantizers.132 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.down_proj_weight_quantizers.133 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.down_proj_weight_quantizers.134 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.down_proj_weight_quantizers.135 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.64e-02 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.down_proj_weight_quantizers.136 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.down_proj_weight_quantizers.137 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.down_proj_weight_quantizers.138 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.down_proj_weight_quantizers.139 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.down_proj_weight_quantizers.140 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.down_proj_weight_quantizers.141 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.down_proj_weight_quantizers.142 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.down_proj_weight_quantizers.143 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.25e-01 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.down_proj_weight_quantizers.144 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.down_proj_weight_quantizers.145 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.23e-01 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.down_proj_weight_quantizers.146 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.down_proj_weight_quantizers.147 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.down_proj_weight_quantizers.148 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.down_proj_weight_quantizers.149 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.18e-01 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.down_proj_weight_quantizers.150 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.64e-02 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.down_proj_weight_quantizers.151 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.19e-01 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.down_proj_weight_quantizers.152 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.down_proj_weight_quantizers.153 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.down_proj_weight_quantizers.154 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.down_proj_weight_quantizers.155 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.21e-01 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.down_proj_weight_quantizers.156 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.down_proj_weight_quantizers.157 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.down_proj_weight_quantizers.158 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.down_proj_weight_quantizers.159 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.30e-01 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.down_proj_weight_quantizers.160 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.33e-01 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.down_proj_weight_quantizers.161 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.down_proj_weight_quantizers.162 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.down_proj_weight_quantizers.163 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.down_proj_weight_quantizers.164 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.64e-02 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.down_proj_weight_quantizers.165 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.16e-01 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.down_proj_weight_quantizers.166 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.down_proj_weight_quantizers.167 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.down_proj_weight_quantizers.168 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.down_proj_weight_quantizers.169 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.87e-01 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.down_proj_weight_quantizers.170 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.down_proj_weight_quantizers.171 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.40e-01 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.down_proj_weight_quantizers.172 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.down_proj_weight_quantizers.173 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.down_proj_weight_quantizers.174 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.down_proj_weight_quantizers.175 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.down_proj_weight_quantizers.176 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.down_proj_weight_quantizers.177 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.60e-01 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.down_proj_weight_quantizers.178 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.down_proj_weight_quantizers.179 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.down_proj_weight_quantizers.180 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.down_proj_weight_quantizers.181 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.down_proj_weight_quantizers.182 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.down_proj_weight_quantizers.183 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.14e-01 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.down_proj_weight_quantizers.184 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.down_proj_weight_quantizers.185 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.down_proj_weight_quantizers.186 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.07e-01 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.down_proj_weight_quantizers.187 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.down_proj_weight_quantizers.188 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.down_proj_weight_quantizers.189 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.64e-02 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.down_proj_weight_quantizers.190 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.21e-01 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.down_proj_weight_quantizers.191 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.78e-01 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.down_proj_weight_quantizers.192 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.down_proj_weight_quantizers.193 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.down_proj_weight_quantizers.194 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.55e-01 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.down_proj_weight_quantizers.195 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.20e-01 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.down_proj_weight_quantizers.196 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.down_proj_weight_quantizers.197 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.22e-01 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.down_proj_weight_quantizers.198 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.down_proj_weight_quantizers.199 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.down_proj_weight_quantizers.200 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.down_proj_weight_quantizers.201 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.down_proj_weight_quantizers.202 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.down_proj_weight_quantizers.203 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.down_proj_weight_quantizers.204 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.21e-01 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.down_proj_weight_quantizers.205 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.down_proj_weight_quantizers.206 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.58e-01 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.down_proj_weight_quantizers.207 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.64e-02 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.down_proj_weight_quantizers.208 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.down_proj_weight_quantizers.209 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.down_proj_weight_quantizers.210 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.down_proj_weight_quantizers.211 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.down_proj_weight_quantizers.212 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.down_proj_weight_quantizers.213 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.down_proj_weight_quantizers.214 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.down_proj_weight_quantizers.215 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.down_proj_weight_quantizers.216 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.down_proj_weight_quantizers.217 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.26e-01 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.down_proj_weight_quantizers.218 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.down_proj_weight_quantizers.219 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.14e-01 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.down_proj_weight_quantizers.220 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.34e-01 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.down_proj_weight_quantizers.221 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.57e-01 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.down_proj_weight_quantizers.222 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.54e-02 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.down_proj_weight_quantizers.223 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.16e-01 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.down_proj_weight_quantizers.224 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.down_proj_weight_quantizers.225 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.down_proj_weight_quantizers.226 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.down_proj_weight_quantizers.227 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.26e-01 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.down_proj_weight_quantizers.228 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.down_proj_weight_quantizers.229 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.down_proj_weight_quantizers.230 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.down_proj_weight_quantizers.231 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.down_proj_weight_quantizers.232 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.down_proj_weight_quantizers.233 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.down_proj_weight_quantizers.234 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.54e-01 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.down_proj_weight_quantizers.235 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.down_proj_weight_quantizers.236 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.down_proj_weight_quantizers.237 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.down_proj_weight_quantizers.238 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.down_proj_weight_quantizers.239 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.down_proj_weight_quantizers.240 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.down_proj_weight_quantizers.241 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.down_proj_weight_quantizers.242 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.61e-01 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.down_proj_weight_quantizers.243 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.down_proj_weight_quantizers.244 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.down_proj_weight_quantizers.245 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.42e-01 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.down_proj_weight_quantizers.246 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.39e-01 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.down_proj_weight_quantizers.247 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.18e-01 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.down_proj_weight_quantizers.248 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.down_proj_weight_quantizers.249 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.down_proj_weight_quantizers.250 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.down_proj_weight_quantizers.251 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.down_proj_weight_quantizers.252 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.down_proj_weight_quantizers.253 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.26e-01 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.down_proj_weight_quantizers.254 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.down_proj_weight_quantizers.255 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.14.mlp.shared_experts.gate_proj.input_quantizer TensorQuantizer(disabled) +model.layers.14.mlp.shared_experts.gate_proj.output_quantizer TensorQuantizer(disabled) +model.layers.14.mlp.shared_experts.gate_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.14.mlp.shared_experts.up_proj.input_quantizer TensorQuantizer(disabled) +model.layers.14.mlp.shared_experts.up_proj.output_quantizer TensorQuantizer(disabled) +model.layers.14.mlp.shared_experts.up_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.14.mlp.shared_experts.down_proj.input_quantizer TensorQuantizer(disabled) +model.layers.14.mlp.shared_experts.down_proj.output_quantizer TensorQuantizer(disabled) +model.layers.14.mlp.shared_experts.down_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.15.self_attn.q_a_proj.input_quantizer TensorQuantizer(disabled) +model.layers.15.self_attn.q_a_proj.output_quantizer TensorQuantizer(disabled) +model.layers.15.self_attn.q_a_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.15.self_attn.q_b_proj.input_quantizer TensorQuantizer(disabled) +model.layers.15.self_attn.q_b_proj.output_quantizer TensorQuantizer(disabled) +model.layers.15.self_attn.q_b_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.15.self_attn.kv_a_proj_with_mqa.input_quantizer TensorQuantizer(disabled) +model.layers.15.self_attn.kv_a_proj_with_mqa.output_quantizer TensorQuantizer(disabled) +model.layers.15.self_attn.kv_a_proj_with_mqa.weight_quantizer TensorQuantizer(disabled) +model.layers.15.self_attn.kv_b_proj.input_quantizer TensorQuantizer(disabled) +model.layers.15.self_attn.kv_b_proj.output_quantizer TensorQuantizer(disabled) +model.layers.15.self_attn.kv_b_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.15.self_attn.o_proj.input_quantizer TensorQuantizer(disabled) +model.layers.15.self_attn.o_proj.output_quantizer TensorQuantizer(disabled) +model.layers.15.self_attn.o_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.15.self_attn.q_bmm_quantizer TensorQuantizer(disabled) +model.layers.15.self_attn.k_bmm_quantizer TensorQuantizer((4, 3) bit fake per-tensor amax=4.48e+02(const) calibrator=MaxCalibrator quant) +model.layers.15.self_attn.v_bmm_quantizer TensorQuantizer((4, 3) bit fake per-tensor amax=4.48e+02(const) calibrator=MaxCalibrator quant) +model.layers.15.self_attn.p_bmm_quantizer TensorQuantizer(disabled) +model.layers.15.mlp.experts.gate_up_proj_input_quantizer TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=5.70e-01 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.gate_up_proj_weight_quantizers.0 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.gate_up_proj_weight_quantizers.1 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.gate_up_proj_weight_quantizers.2 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.gate_up_proj_weight_quantizers.3 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.gate_up_proj_weight_quantizers.4 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.gate_up_proj_weight_quantizers.5 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.77e-01 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.gate_up_proj_weight_quantizers.6 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.gate_up_proj_weight_quantizers.7 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.gate_up_proj_weight_quantizers.8 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.gate_up_proj_weight_quantizers.9 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.26e-01 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.gate_up_proj_weight_quantizers.10 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.gate_up_proj_weight_quantizers.11 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.61e-01 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.gate_up_proj_weight_quantizers.12 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.gate_up_proj_weight_quantizers.13 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.gate_up_proj_weight_quantizers.14 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.gate_up_proj_weight_quantizers.15 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.71e-01 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.gate_up_proj_weight_quantizers.16 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.gate_up_proj_weight_quantizers.17 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.gate_up_proj_weight_quantizers.18 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.gate_up_proj_weight_quantizers.19 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.gate_up_proj_weight_quantizers.20 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.gate_up_proj_weight_quantizers.21 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.gate_up_proj_weight_quantizers.22 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.gate_up_proj_weight_quantizers.23 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.52e-01 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.gate_up_proj_weight_quantizers.24 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.gate_up_proj_weight_quantizers.25 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.gate_up_proj_weight_quantizers.26 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.gate_up_proj_weight_quantizers.27 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.gate_up_proj_weight_quantizers.28 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.49e-01 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.gate_up_proj_weight_quantizers.29 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.gate_up_proj_weight_quantizers.30 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.gate_up_proj_weight_quantizers.31 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.gate_up_proj_weight_quantizers.32 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.gate_up_proj_weight_quantizers.33 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.25e-01 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.gate_up_proj_weight_quantizers.34 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.gate_up_proj_weight_quantizers.35 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.14e-01 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.gate_up_proj_weight_quantizers.36 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.gate_up_proj_weight_quantizers.37 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.45e-01 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.gate_up_proj_weight_quantizers.38 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.50e-01 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.gate_up_proj_weight_quantizers.39 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.gate_up_proj_weight_quantizers.40 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.gate_up_proj_weight_quantizers.41 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.32e-01 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.gate_up_proj_weight_quantizers.42 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.gate_up_proj_weight_quantizers.43 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.95e-01 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.gate_up_proj_weight_quantizers.44 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.gate_up_proj_weight_quantizers.45 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.gate_up_proj_weight_quantizers.46 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.26e-01 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.gate_up_proj_weight_quantizers.47 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.gate_up_proj_weight_quantizers.48 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.16e-01 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.gate_up_proj_weight_quantizers.49 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.42e-01 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.gate_up_proj_weight_quantizers.50 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.gate_up_proj_weight_quantizers.51 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.gate_up_proj_weight_quantizers.52 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.gate_up_proj_weight_quantizers.53 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.gate_up_proj_weight_quantizers.54 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.89e-01 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.gate_up_proj_weight_quantizers.55 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.gate_up_proj_weight_quantizers.56 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.gate_up_proj_weight_quantizers.57 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.gate_up_proj_weight_quantizers.58 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.gate_up_proj_weight_quantizers.59 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.gate_up_proj_weight_quantizers.60 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.gate_up_proj_weight_quantizers.61 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.gate_up_proj_weight_quantizers.62 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.gate_up_proj_weight_quantizers.63 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.gate_up_proj_weight_quantizers.64 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.gate_up_proj_weight_quantizers.65 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.gate_up_proj_weight_quantizers.66 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.gate_up_proj_weight_quantizers.67 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.gate_up_proj_weight_quantizers.68 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.gate_up_proj_weight_quantizers.69 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.gate_up_proj_weight_quantizers.70 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.gate_up_proj_weight_quantizers.71 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.gate_up_proj_weight_quantizers.72 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.gate_up_proj_weight_quantizers.73 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.gate_up_proj_weight_quantizers.74 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.gate_up_proj_weight_quantizers.75 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.gate_up_proj_weight_quantizers.76 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.69e-01 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.gate_up_proj_weight_quantizers.77 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.gate_up_proj_weight_quantizers.78 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.30e-01 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.gate_up_proj_weight_quantizers.79 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.gate_up_proj_weight_quantizers.80 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.54e-01 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.gate_up_proj_weight_quantizers.81 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.gate_up_proj_weight_quantizers.82 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.gate_up_proj_weight_quantizers.83 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.gate_up_proj_weight_quantizers.84 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.43e-01 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.gate_up_proj_weight_quantizers.85 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.gate_up_proj_weight_quantizers.86 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.gate_up_proj_weight_quantizers.87 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.gate_up_proj_weight_quantizers.88 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.gate_up_proj_weight_quantizers.89 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.gate_up_proj_weight_quantizers.90 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.gate_up_proj_weight_quantizers.91 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.gate_up_proj_weight_quantizers.92 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.31e-01 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.gate_up_proj_weight_quantizers.93 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.gate_up_proj_weight_quantizers.94 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.gate_up_proj_weight_quantizers.95 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.gate_up_proj_weight_quantizers.96 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.gate_up_proj_weight_quantizers.97 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.gate_up_proj_weight_quantizers.98 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.gate_up_proj_weight_quantizers.99 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.gate_up_proj_weight_quantizers.100 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.05e-01 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.gate_up_proj_weight_quantizers.101 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.gate_up_proj_weight_quantizers.102 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.23e-01 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.gate_up_proj_weight_quantizers.103 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.gate_up_proj_weight_quantizers.104 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.gate_up_proj_weight_quantizers.105 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.gate_up_proj_weight_quantizers.106 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.gate_up_proj_weight_quantizers.107 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.gate_up_proj_weight_quantizers.108 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.gate_up_proj_weight_quantizers.109 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.20e-01 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.gate_up_proj_weight_quantizers.110 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.22e-01 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.gate_up_proj_weight_quantizers.111 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.gate_up_proj_weight_quantizers.112 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.gate_up_proj_weight_quantizers.113 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.gate_up_proj_weight_quantizers.114 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.21e-01 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.gate_up_proj_weight_quantizers.115 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.gate_up_proj_weight_quantizers.116 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.gate_up_proj_weight_quantizers.117 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.gate_up_proj_weight_quantizers.118 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.gate_up_proj_weight_quantizers.119 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.gate_up_proj_weight_quantizers.120 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.20e-01 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.gate_up_proj_weight_quantizers.121 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.gate_up_proj_weight_quantizers.122 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.gate_up_proj_weight_quantizers.123 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.18e-01 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.gate_up_proj_weight_quantizers.124 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.gate_up_proj_weight_quantizers.125 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.gate_up_proj_weight_quantizers.126 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.gate_up_proj_weight_quantizers.127 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.65e-01 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.gate_up_proj_weight_quantizers.128 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.gate_up_proj_weight_quantizers.129 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.gate_up_proj_weight_quantizers.130 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.gate_up_proj_weight_quantizers.131 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.gate_up_proj_weight_quantizers.132 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.gate_up_proj_weight_quantizers.133 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.gate_up_proj_weight_quantizers.134 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.gate_up_proj_weight_quantizers.135 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.gate_up_proj_weight_quantizers.136 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.gate_up_proj_weight_quantizers.137 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.75e-01 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.gate_up_proj_weight_quantizers.138 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.gate_up_proj_weight_quantizers.139 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.gate_up_proj_weight_quantizers.140 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.gate_up_proj_weight_quantizers.141 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=7.73e-01 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.gate_up_proj_weight_quantizers.142 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.gate_up_proj_weight_quantizers.143 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.56e-01 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.gate_up_proj_weight_quantizers.144 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.29e-01 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.gate_up_proj_weight_quantizers.145 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.22e-01 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.gate_up_proj_weight_quantizers.146 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.33e-01 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.gate_up_proj_weight_quantizers.147 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.gate_up_proj_weight_quantizers.148 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.gate_up_proj_weight_quantizers.149 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.gate_up_proj_weight_quantizers.150 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.gate_up_proj_weight_quantizers.151 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.gate_up_proj_weight_quantizers.152 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.gate_up_proj_weight_quantizers.153 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.gate_up_proj_weight_quantizers.154 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.74e-01 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.gate_up_proj_weight_quantizers.155 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.gate_up_proj_weight_quantizers.156 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.gate_up_proj_weight_quantizers.157 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.19e-01 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.gate_up_proj_weight_quantizers.158 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.gate_up_proj_weight_quantizers.159 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.72e-01 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.gate_up_proj_weight_quantizers.160 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.gate_up_proj_weight_quantizers.161 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.gate_up_proj_weight_quantizers.162 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.gate_up_proj_weight_quantizers.163 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.gate_up_proj_weight_quantizers.164 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.gate_up_proj_weight_quantizers.165 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.gate_up_proj_weight_quantizers.166 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.gate_up_proj_weight_quantizers.167 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.gate_up_proj_weight_quantizers.168 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.gate_up_proj_weight_quantizers.169 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.gate_up_proj_weight_quantizers.170 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.gate_up_proj_weight_quantizers.171 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.gate_up_proj_weight_quantizers.172 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.gate_up_proj_weight_quantizers.173 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.gate_up_proj_weight_quantizers.174 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.gate_up_proj_weight_quantizers.175 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.gate_up_proj_weight_quantizers.176 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.gate_up_proj_weight_quantizers.177 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.gate_up_proj_weight_quantizers.178 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.gate_up_proj_weight_quantizers.179 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.gate_up_proj_weight_quantizers.180 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.gate_up_proj_weight_quantizers.181 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.52e-01 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.gate_up_proj_weight_quantizers.182 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.gate_up_proj_weight_quantizers.183 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.gate_up_proj_weight_quantizers.184 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.24e-01 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.gate_up_proj_weight_quantizers.185 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.22e-01 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.gate_up_proj_weight_quantizers.186 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.gate_up_proj_weight_quantizers.187 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.56e-01 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.gate_up_proj_weight_quantizers.188 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.gate_up_proj_weight_quantizers.189 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.gate_up_proj_weight_quantizers.190 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.gate_up_proj_weight_quantizers.191 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.gate_up_proj_weight_quantizers.192 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.gate_up_proj_weight_quantizers.193 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.gate_up_proj_weight_quantizers.194 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.gate_up_proj_weight_quantizers.195 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.gate_up_proj_weight_quantizers.196 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.gate_up_proj_weight_quantizers.197 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.21e-01 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.gate_up_proj_weight_quantizers.198 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.gate_up_proj_weight_quantizers.199 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.gate_up_proj_weight_quantizers.200 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.gate_up_proj_weight_quantizers.201 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.78e-01 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.gate_up_proj_weight_quantizers.202 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.gate_up_proj_weight_quantizers.203 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.gate_up_proj_weight_quantizers.204 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.gate_up_proj_weight_quantizers.205 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.gate_up_proj_weight_quantizers.206 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.gate_up_proj_weight_quantizers.207 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.gate_up_proj_weight_quantizers.208 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.gate_up_proj_weight_quantizers.209 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.gate_up_proj_weight_quantizers.210 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.gate_up_proj_weight_quantizers.211 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.gate_up_proj_weight_quantizers.212 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.gate_up_proj_weight_quantizers.213 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.gate_up_proj_weight_quantizers.214 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.gate_up_proj_weight_quantizers.215 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.gate_up_proj_weight_quantizers.216 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.gate_up_proj_weight_quantizers.217 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.gate_up_proj_weight_quantizers.218 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.gate_up_proj_weight_quantizers.219 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.71e-01 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.gate_up_proj_weight_quantizers.220 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.gate_up_proj_weight_quantizers.221 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.gate_up_proj_weight_quantizers.222 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.73e-01 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.gate_up_proj_weight_quantizers.223 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.gate_up_proj_weight_quantizers.224 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.gate_up_proj_weight_quantizers.225 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.gate_up_proj_weight_quantizers.226 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.gate_up_proj_weight_quantizers.227 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.gate_up_proj_weight_quantizers.228 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.gate_up_proj_weight_quantizers.229 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.gate_up_proj_weight_quantizers.230 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.gate_up_proj_weight_quantizers.231 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.15e-01 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.gate_up_proj_weight_quantizers.232 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.gate_up_proj_weight_quantizers.233 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.gate_up_proj_weight_quantizers.234 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=5.94e-01 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.gate_up_proj_weight_quantizers.235 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.gate_up_proj_weight_quantizers.236 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.gate_up_proj_weight_quantizers.237 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.gate_up_proj_weight_quantizers.238 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.gate_up_proj_weight_quantizers.239 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.85e-01 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.gate_up_proj_weight_quantizers.240 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.gate_up_proj_weight_quantizers.241 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.gate_up_proj_weight_quantizers.242 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.gate_up_proj_weight_quantizers.243 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.42e-01 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.gate_up_proj_weight_quantizers.244 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.gate_up_proj_weight_quantizers.245 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.95e-01 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.gate_up_proj_weight_quantizers.246 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.gate_up_proj_weight_quantizers.247 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.gate_up_proj_weight_quantizers.248 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.gate_up_proj_weight_quantizers.249 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.33e-01 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.gate_up_proj_weight_quantizers.250 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.88e-01 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.gate_up_proj_weight_quantizers.251 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.gate_up_proj_weight_quantizers.252 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.gate_up_proj_weight_quantizers.253 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.gate_up_proj_weight_quantizers.254 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.gate_up_proj_weight_quantizers.255 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.down_proj_input_quantizer TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.25e+00 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.down_proj_weight_quantizers.0 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.down_proj_weight_quantizers.1 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.64e-02 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.down_proj_weight_quantizers.2 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.down_proj_weight_quantizers.3 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.down_proj_weight_quantizers.4 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.down_proj_weight_quantizers.5 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.48e-01 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.down_proj_weight_quantizers.6 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.down_proj_weight_quantizers.7 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.down_proj_weight_quantizers.8 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.down_proj_weight_quantizers.9 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.24e-01 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.down_proj_weight_quantizers.10 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.down_proj_weight_quantizers.11 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.down_proj_weight_quantizers.12 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.down_proj_weight_quantizers.13 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.down_proj_weight_quantizers.14 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.down_proj_weight_quantizers.15 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.58e-01 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.down_proj_weight_quantizers.16 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.13e-01 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.down_proj_weight_quantizers.17 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.64e-02 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.down_proj_weight_quantizers.18 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.41e-01 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.down_proj_weight_quantizers.19 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.18e-01 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.down_proj_weight_quantizers.20 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.down_proj_weight_quantizers.21 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.down_proj_weight_quantizers.22 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.down_proj_weight_quantizers.23 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.97e-01 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.down_proj_weight_quantizers.24 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.down_proj_weight_quantizers.25 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.down_proj_weight_quantizers.26 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.down_proj_weight_quantizers.27 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.down_proj_weight_quantizers.28 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.49e-01 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.down_proj_weight_quantizers.29 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.down_proj_weight_quantizers.30 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.down_proj_weight_quantizers.31 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.33e-01 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.down_proj_weight_quantizers.32 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.down_proj_weight_quantizers.33 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.down_proj_weight_quantizers.34 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.20e-01 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.down_proj_weight_quantizers.35 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.31e-01 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.down_proj_weight_quantizers.36 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.down_proj_weight_quantizers.37 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.30e-01 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.down_proj_weight_quantizers.38 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.down_proj_weight_quantizers.39 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.down_proj_weight_quantizers.40 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.54e-01 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.down_proj_weight_quantizers.41 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.down_proj_weight_quantizers.42 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.down_proj_weight_quantizers.43 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.73e-01 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.down_proj_weight_quantizers.44 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.down_proj_weight_quantizers.45 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.down_proj_weight_quantizers.46 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.down_proj_weight_quantizers.47 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.down_proj_weight_quantizers.48 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.17e-01 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.down_proj_weight_quantizers.49 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.32e-01 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.down_proj_weight_quantizers.50 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.down_proj_weight_quantizers.51 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.down_proj_weight_quantizers.52 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.down_proj_weight_quantizers.53 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.down_proj_weight_quantizers.54 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.down_proj_weight_quantizers.55 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.down_proj_weight_quantizers.56 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.down_proj_weight_quantizers.57 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.down_proj_weight_quantizers.58 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.64e-02 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.down_proj_weight_quantizers.59 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.down_proj_weight_quantizers.60 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.40e-02 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.down_proj_weight_quantizers.61 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.down_proj_weight_quantizers.62 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.down_proj_weight_quantizers.63 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.down_proj_weight_quantizers.64 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.down_proj_weight_quantizers.65 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.down_proj_weight_quantizers.66 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.15e-01 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.down_proj_weight_quantizers.67 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.down_proj_weight_quantizers.68 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.down_proj_weight_quantizers.69 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.down_proj_weight_quantizers.70 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.59e-02 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.down_proj_weight_quantizers.71 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.down_proj_weight_quantizers.72 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.down_proj_weight_quantizers.73 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.down_proj_weight_quantizers.74 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.down_proj_weight_quantizers.75 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.23e-01 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.down_proj_weight_quantizers.76 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.43e-01 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.down_proj_weight_quantizers.77 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.down_proj_weight_quantizers.78 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.73e-01 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.down_proj_weight_quantizers.79 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.down_proj_weight_quantizers.80 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.down_proj_weight_quantizers.81 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.down_proj_weight_quantizers.82 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.down_proj_weight_quantizers.83 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.down_proj_weight_quantizers.84 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.75e-01 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.down_proj_weight_quantizers.85 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.down_proj_weight_quantizers.86 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.50e-01 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.down_proj_weight_quantizers.87 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.down_proj_weight_quantizers.88 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.down_proj_weight_quantizers.89 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.down_proj_weight_quantizers.90 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.down_proj_weight_quantizers.91 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.down_proj_weight_quantizers.92 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.down_proj_weight_quantizers.93 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.down_proj_weight_quantizers.94 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.down_proj_weight_quantizers.95 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.83e-01 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.down_proj_weight_quantizers.96 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.64e-02 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.down_proj_weight_quantizers.97 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.down_proj_weight_quantizers.98 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.down_proj_weight_quantizers.99 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.down_proj_weight_quantizers.100 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.57e-01 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.down_proj_weight_quantizers.101 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.down_proj_weight_quantizers.102 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.15e-01 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.down_proj_weight_quantizers.103 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.down_proj_weight_quantizers.104 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.down_proj_weight_quantizers.105 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.down_proj_weight_quantizers.106 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.down_proj_weight_quantizers.107 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.down_proj_weight_quantizers.108 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.down_proj_weight_quantizers.109 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.28e-01 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.down_proj_weight_quantizers.110 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.down_proj_weight_quantizers.111 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.down_proj_weight_quantizers.112 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.down_proj_weight_quantizers.113 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.25e-01 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.down_proj_weight_quantizers.114 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.down_proj_weight_quantizers.115 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.56e-01 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.down_proj_weight_quantizers.116 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.down_proj_weight_quantizers.117 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.33e-01 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.down_proj_weight_quantizers.118 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.down_proj_weight_quantizers.119 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.down_proj_weight_quantizers.120 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.down_proj_weight_quantizers.121 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.25e-01 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.down_proj_weight_quantizers.122 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.down_proj_weight_quantizers.123 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.down_proj_weight_quantizers.124 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.65e-01 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.down_proj_weight_quantizers.125 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.down_proj_weight_quantizers.126 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.46e-01 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.down_proj_weight_quantizers.127 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.down_proj_weight_quantizers.128 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.down_proj_weight_quantizers.129 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.down_proj_weight_quantizers.130 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.down_proj_weight_quantizers.131 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.down_proj_weight_quantizers.132 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.down_proj_weight_quantizers.133 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.down_proj_weight_quantizers.134 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.21e-01 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.down_proj_weight_quantizers.135 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.down_proj_weight_quantizers.136 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.down_proj_weight_quantizers.137 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.down_proj_weight_quantizers.138 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.down_proj_weight_quantizers.139 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.16e-01 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.down_proj_weight_quantizers.140 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.down_proj_weight_quantizers.141 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.48e-01 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.down_proj_weight_quantizers.142 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.down_proj_weight_quantizers.143 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.down_proj_weight_quantizers.144 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.down_proj_weight_quantizers.145 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.65e-01 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.down_proj_weight_quantizers.146 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.47e-01 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.down_proj_weight_quantizers.147 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.down_proj_weight_quantizers.148 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.18e-01 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.down_proj_weight_quantizers.149 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.down_proj_weight_quantizers.150 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.down_proj_weight_quantizers.151 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.down_proj_weight_quantizers.152 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.down_proj_weight_quantizers.153 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.down_proj_weight_quantizers.154 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.down_proj_weight_quantizers.155 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.20e-01 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.down_proj_weight_quantizers.156 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.down_proj_weight_quantizers.157 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.15e-01 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.down_proj_weight_quantizers.158 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.down_proj_weight_quantizers.159 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.51e-01 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.down_proj_weight_quantizers.160 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.24e-01 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.down_proj_weight_quantizers.161 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.down_proj_weight_quantizers.162 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.down_proj_weight_quantizers.163 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.50e-02 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.down_proj_weight_quantizers.164 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.down_proj_weight_quantizers.165 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.down_proj_weight_quantizers.166 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.86e-01 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.down_proj_weight_quantizers.167 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.down_proj_weight_quantizers.168 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.down_proj_weight_quantizers.169 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.down_proj_weight_quantizers.170 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.down_proj_weight_quantizers.171 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.down_proj_weight_quantizers.172 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.13e-01 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.down_proj_weight_quantizers.173 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.down_proj_weight_quantizers.174 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.down_proj_weight_quantizers.175 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.35e-01 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.down_proj_weight_quantizers.176 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.down_proj_weight_quantizers.177 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.down_proj_weight_quantizers.178 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.down_proj_weight_quantizers.179 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.down_proj_weight_quantizers.180 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.down_proj_weight_quantizers.181 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.down_proj_weight_quantizers.182 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.down_proj_weight_quantizers.183 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.down_proj_weight_quantizers.184 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.31e-01 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.down_proj_weight_quantizers.185 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.down_proj_weight_quantizers.186 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.down_proj_weight_quantizers.187 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.80e-01 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.down_proj_weight_quantizers.188 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.42e-01 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.down_proj_weight_quantizers.189 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.14e-01 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.down_proj_weight_quantizers.190 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.down_proj_weight_quantizers.191 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.down_proj_weight_quantizers.192 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.down_proj_weight_quantizers.193 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.down_proj_weight_quantizers.194 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.down_proj_weight_quantizers.195 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.down_proj_weight_quantizers.196 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.down_proj_weight_quantizers.197 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.down_proj_weight_quantizers.198 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.down_proj_weight_quantizers.199 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.down_proj_weight_quantizers.200 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.down_proj_weight_quantizers.201 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.35e-01 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.down_proj_weight_quantizers.202 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.44e-01 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.down_proj_weight_quantizers.203 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.13e-01 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.down_proj_weight_quantizers.204 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.33e-01 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.down_proj_weight_quantizers.205 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.down_proj_weight_quantizers.206 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.down_proj_weight_quantizers.207 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.down_proj_weight_quantizers.208 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.down_proj_weight_quantizers.209 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.down_proj_weight_quantizers.210 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.22e-01 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.down_proj_weight_quantizers.211 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.down_proj_weight_quantizers.212 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.down_proj_weight_quantizers.213 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.down_proj_weight_quantizers.214 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.down_proj_weight_quantizers.215 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.down_proj_weight_quantizers.216 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.31e-01 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.down_proj_weight_quantizers.217 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.down_proj_weight_quantizers.218 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.64e-02 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.down_proj_weight_quantizers.219 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.67e-01 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.down_proj_weight_quantizers.220 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.21e-01 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.down_proj_weight_quantizers.221 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.down_proj_weight_quantizers.222 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=5.86e-01 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.down_proj_weight_quantizers.223 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.down_proj_weight_quantizers.224 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.49e-01 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.down_proj_weight_quantizers.225 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.down_proj_weight_quantizers.226 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.down_proj_weight_quantizers.227 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.down_proj_weight_quantizers.228 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.down_proj_weight_quantizers.229 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.down_proj_weight_quantizers.230 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.15e-01 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.down_proj_weight_quantizers.231 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.down_proj_weight_quantizers.232 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.down_proj_weight_quantizers.233 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.16e-01 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.down_proj_weight_quantizers.234 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.24e-01 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.down_proj_weight_quantizers.235 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.down_proj_weight_quantizers.236 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.15e-01 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.down_proj_weight_quantizers.237 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.down_proj_weight_quantizers.238 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.32e-01 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.down_proj_weight_quantizers.239 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.63e-01 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.down_proj_weight_quantizers.240 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.down_proj_weight_quantizers.241 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.down_proj_weight_quantizers.242 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.down_proj_weight_quantizers.243 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.down_proj_weight_quantizers.244 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.down_proj_weight_quantizers.245 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.82e-01 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.down_proj_weight_quantizers.246 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.39e-01 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.down_proj_weight_quantizers.247 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.down_proj_weight_quantizers.248 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.50e-02 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.down_proj_weight_quantizers.249 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.21e-01 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.down_proj_weight_quantizers.250 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.18e-01 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.down_proj_weight_quantizers.251 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.down_proj_weight_quantizers.252 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.down_proj_weight_quantizers.253 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.59e-02 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.down_proj_weight_quantizers.254 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.down_proj_weight_quantizers.255 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.15.mlp.shared_experts.gate_proj.input_quantizer TensorQuantizer(disabled) +model.layers.15.mlp.shared_experts.gate_proj.output_quantizer TensorQuantizer(disabled) +model.layers.15.mlp.shared_experts.gate_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.15.mlp.shared_experts.up_proj.input_quantizer TensorQuantizer(disabled) +model.layers.15.mlp.shared_experts.up_proj.output_quantizer TensorQuantizer(disabled) +model.layers.15.mlp.shared_experts.up_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.15.mlp.shared_experts.down_proj.input_quantizer TensorQuantizer(disabled) +model.layers.15.mlp.shared_experts.down_proj.output_quantizer TensorQuantizer(disabled) +model.layers.15.mlp.shared_experts.down_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.16.self_attn.q_a_proj.input_quantizer TensorQuantizer(disabled) +model.layers.16.self_attn.q_a_proj.output_quantizer TensorQuantizer(disabled) +model.layers.16.self_attn.q_a_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.16.self_attn.q_b_proj.input_quantizer TensorQuantizer(disabled) +model.layers.16.self_attn.q_b_proj.output_quantizer TensorQuantizer(disabled) +model.layers.16.self_attn.q_b_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.16.self_attn.kv_a_proj_with_mqa.input_quantizer TensorQuantizer(disabled) +model.layers.16.self_attn.kv_a_proj_with_mqa.output_quantizer TensorQuantizer(disabled) +model.layers.16.self_attn.kv_a_proj_with_mqa.weight_quantizer TensorQuantizer(disabled) +model.layers.16.self_attn.kv_b_proj.input_quantizer TensorQuantizer(disabled) +model.layers.16.self_attn.kv_b_proj.output_quantizer TensorQuantizer(disabled) +model.layers.16.self_attn.kv_b_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.16.self_attn.o_proj.input_quantizer TensorQuantizer(disabled) +model.layers.16.self_attn.o_proj.output_quantizer TensorQuantizer(disabled) +model.layers.16.self_attn.o_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.16.self_attn.q_bmm_quantizer TensorQuantizer(disabled) +model.layers.16.self_attn.k_bmm_quantizer TensorQuantizer((4, 3) bit fake per-tensor amax=4.48e+02(const) calibrator=MaxCalibrator quant) +model.layers.16.self_attn.v_bmm_quantizer TensorQuantizer((4, 3) bit fake per-tensor amax=4.48e+02(const) calibrator=MaxCalibrator quant) +model.layers.16.self_attn.p_bmm_quantizer TensorQuantizer(disabled) +model.layers.16.mlp.experts.gate_up_proj_input_quantizer TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=5.66e-01 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.gate_up_proj_weight_quantizers.0 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.gate_up_proj_weight_quantizers.1 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.gate_up_proj_weight_quantizers.2 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.gate_up_proj_weight_quantizers.3 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.28e-01 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.gate_up_proj_weight_quantizers.4 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.gate_up_proj_weight_quantizers.5 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.gate_up_proj_weight_quantizers.6 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.31e-01 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.gate_up_proj_weight_quantizers.7 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.gate_up_proj_weight_quantizers.8 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.gate_up_proj_weight_quantizers.9 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.gate_up_proj_weight_quantizers.10 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.gate_up_proj_weight_quantizers.11 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.gate_up_proj_weight_quantizers.12 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.gate_up_proj_weight_quantizers.13 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.gate_up_proj_weight_quantizers.14 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.35e-01 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.gate_up_proj_weight_quantizers.15 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.gate_up_proj_weight_quantizers.16 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.gate_up_proj_weight_quantizers.17 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.gate_up_proj_weight_quantizers.18 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.gate_up_proj_weight_quantizers.19 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.gate_up_proj_weight_quantizers.20 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.gate_up_proj_weight_quantizers.21 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.gate_up_proj_weight_quantizers.22 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.gate_up_proj_weight_quantizers.23 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.gate_up_proj_weight_quantizers.24 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.gate_up_proj_weight_quantizers.25 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.20e-01 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.gate_up_proj_weight_quantizers.26 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.45e-01 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.gate_up_proj_weight_quantizers.27 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.gate_up_proj_weight_quantizers.28 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.gate_up_proj_weight_quantizers.29 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.99e-01 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.gate_up_proj_weight_quantizers.30 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.gate_up_proj_weight_quantizers.31 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.gate_up_proj_weight_quantizers.32 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.gate_up_proj_weight_quantizers.33 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.gate_up_proj_weight_quantizers.34 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.35e-01 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.gate_up_proj_weight_quantizers.35 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.66e-01 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.gate_up_proj_weight_quantizers.36 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.gate_up_proj_weight_quantizers.37 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.gate_up_proj_weight_quantizers.38 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.50e-01 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.gate_up_proj_weight_quantizers.39 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.gate_up_proj_weight_quantizers.40 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.gate_up_proj_weight_quantizers.41 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.gate_up_proj_weight_quantizers.42 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.gate_up_proj_weight_quantizers.43 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.43e-01 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.gate_up_proj_weight_quantizers.44 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.gate_up_proj_weight_quantizers.45 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.34e-01 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.gate_up_proj_weight_quantizers.46 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.41e-01 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.gate_up_proj_weight_quantizers.47 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.gate_up_proj_weight_quantizers.48 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.gate_up_proj_weight_quantizers.49 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.30e-01 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.gate_up_proj_weight_quantizers.50 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.45e-01 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.gate_up_proj_weight_quantizers.51 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.gate_up_proj_weight_quantizers.52 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.gate_up_proj_weight_quantizers.53 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.33e-01 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.gate_up_proj_weight_quantizers.54 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.gate_up_proj_weight_quantizers.55 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.gate_up_proj_weight_quantizers.56 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.gate_up_proj_weight_quantizers.57 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.gate_up_proj_weight_quantizers.58 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.gate_up_proj_weight_quantizers.59 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.gate_up_proj_weight_quantizers.60 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.33e-01 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.gate_up_proj_weight_quantizers.61 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.gate_up_proj_weight_quantizers.62 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.gate_up_proj_weight_quantizers.63 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.gate_up_proj_weight_quantizers.64 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.gate_up_proj_weight_quantizers.65 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.gate_up_proj_weight_quantizers.66 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.gate_up_proj_weight_quantizers.67 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.gate_up_proj_weight_quantizers.68 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.gate_up_proj_weight_quantizers.69 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.gate_up_proj_weight_quantizers.70 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.gate_up_proj_weight_quantizers.71 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.gate_up_proj_weight_quantizers.72 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.gate_up_proj_weight_quantizers.73 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.36e-01 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.gate_up_proj_weight_quantizers.74 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.29e-01 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.gate_up_proj_weight_quantizers.75 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.gate_up_proj_weight_quantizers.76 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.gate_up_proj_weight_quantizers.77 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.gate_up_proj_weight_quantizers.78 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.30e-01 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.gate_up_proj_weight_quantizers.79 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.gate_up_proj_weight_quantizers.80 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.gate_up_proj_weight_quantizers.81 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.gate_up_proj_weight_quantizers.82 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.18e-01 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.gate_up_proj_weight_quantizers.83 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.gate_up_proj_weight_quantizers.84 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.56e-01 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.gate_up_proj_weight_quantizers.85 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.gate_up_proj_weight_quantizers.86 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.gate_up_proj_weight_quantizers.87 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.gate_up_proj_weight_quantizers.88 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.gate_up_proj_weight_quantizers.89 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.16e-01 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.gate_up_proj_weight_quantizers.90 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.gate_up_proj_weight_quantizers.91 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.gate_up_proj_weight_quantizers.92 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.gate_up_proj_weight_quantizers.93 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.gate_up_proj_weight_quantizers.94 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.gate_up_proj_weight_quantizers.95 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.gate_up_proj_weight_quantizers.96 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.gate_up_proj_weight_quantizers.97 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.gate_up_proj_weight_quantizers.98 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.gate_up_proj_weight_quantizers.99 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.gate_up_proj_weight_quantizers.100 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.gate_up_proj_weight_quantizers.101 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.gate_up_proj_weight_quantizers.102 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.gate_up_proj_weight_quantizers.103 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.gate_up_proj_weight_quantizers.104 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.gate_up_proj_weight_quantizers.105 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.gate_up_proj_weight_quantizers.106 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.22e-01 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.gate_up_proj_weight_quantizers.107 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.gate_up_proj_weight_quantizers.108 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.59e-01 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.gate_up_proj_weight_quantizers.109 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.gate_up_proj_weight_quantizers.110 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.26e-01 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.gate_up_proj_weight_quantizers.111 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.10e-01 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.gate_up_proj_weight_quantizers.112 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.gate_up_proj_weight_quantizers.113 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.11e-01 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.gate_up_proj_weight_quantizers.114 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.gate_up_proj_weight_quantizers.115 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.gate_up_proj_weight_quantizers.116 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.gate_up_proj_weight_quantizers.117 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.gate_up_proj_weight_quantizers.118 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.gate_up_proj_weight_quantizers.119 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.gate_up_proj_weight_quantizers.120 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.gate_up_proj_weight_quantizers.121 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.gate_up_proj_weight_quantizers.122 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.41e-01 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.gate_up_proj_weight_quantizers.123 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.gate_up_proj_weight_quantizers.124 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.13e-01 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.gate_up_proj_weight_quantizers.125 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.gate_up_proj_weight_quantizers.126 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.gate_up_proj_weight_quantizers.127 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.gate_up_proj_weight_quantizers.128 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.gate_up_proj_weight_quantizers.129 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.gate_up_proj_weight_quantizers.130 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.53e-01 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.gate_up_proj_weight_quantizers.131 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.gate_up_proj_weight_quantizers.132 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.gate_up_proj_weight_quantizers.133 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.gate_up_proj_weight_quantizers.134 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.gate_up_proj_weight_quantizers.135 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.27e-01 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.gate_up_proj_weight_quantizers.136 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.17e-01 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.gate_up_proj_weight_quantizers.137 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.gate_up_proj_weight_quantizers.138 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.gate_up_proj_weight_quantizers.139 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.gate_up_proj_weight_quantizers.140 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.gate_up_proj_weight_quantizers.141 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.gate_up_proj_weight_quantizers.142 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.gate_up_proj_weight_quantizers.143 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.gate_up_proj_weight_quantizers.144 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.gate_up_proj_weight_quantizers.145 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.57e-01 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.gate_up_proj_weight_quantizers.146 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.24e-01 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.gate_up_proj_weight_quantizers.147 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.gate_up_proj_weight_quantizers.148 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.gate_up_proj_weight_quantizers.149 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.gate_up_proj_weight_quantizers.150 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.gate_up_proj_weight_quantizers.151 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.gate_up_proj_weight_quantizers.152 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.gate_up_proj_weight_quantizers.153 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.34e-01 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.gate_up_proj_weight_quantizers.154 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.19e-01 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.gate_up_proj_weight_quantizers.155 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.gate_up_proj_weight_quantizers.156 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.gate_up_proj_weight_quantizers.157 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.gate_up_proj_weight_quantizers.158 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.gate_up_proj_weight_quantizers.159 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.17e-01 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.gate_up_proj_weight_quantizers.160 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.14e-01 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.gate_up_proj_weight_quantizers.161 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.gate_up_proj_weight_quantizers.162 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.gate_up_proj_weight_quantizers.163 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.gate_up_proj_weight_quantizers.164 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.gate_up_proj_weight_quantizers.165 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.gate_up_proj_weight_quantizers.166 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.gate_up_proj_weight_quantizers.167 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.gate_up_proj_weight_quantizers.168 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.gate_up_proj_weight_quantizers.169 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.gate_up_proj_weight_quantizers.170 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.gate_up_proj_weight_quantizers.171 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.gate_up_proj_weight_quantizers.172 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.gate_up_proj_weight_quantizers.173 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.55e-01 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.gate_up_proj_weight_quantizers.174 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.gate_up_proj_weight_quantizers.175 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.gate_up_proj_weight_quantizers.176 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.gate_up_proj_weight_quantizers.177 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.34e-01 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.gate_up_proj_weight_quantizers.178 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.gate_up_proj_weight_quantizers.179 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.gate_up_proj_weight_quantizers.180 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.74e-01 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.gate_up_proj_weight_quantizers.181 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.gate_up_proj_weight_quantizers.182 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.gate_up_proj_weight_quantizers.183 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.09e-01 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.gate_up_proj_weight_quantizers.184 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.gate_up_proj_weight_quantizers.185 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.gate_up_proj_weight_quantizers.186 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.05e-01 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.gate_up_proj_weight_quantizers.187 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.99e-01 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.gate_up_proj_weight_quantizers.188 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.gate_up_proj_weight_quantizers.189 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.gate_up_proj_weight_quantizers.190 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.gate_up_proj_weight_quantizers.191 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.gate_up_proj_weight_quantizers.192 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.gate_up_proj_weight_quantizers.193 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.22e-01 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.gate_up_proj_weight_quantizers.194 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.gate_up_proj_weight_quantizers.195 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.gate_up_proj_weight_quantizers.196 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.gate_up_proj_weight_quantizers.197 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.gate_up_proj_weight_quantizers.198 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.23e-01 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.gate_up_proj_weight_quantizers.199 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.gate_up_proj_weight_quantizers.200 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.gate_up_proj_weight_quantizers.201 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.gate_up_proj_weight_quantizers.202 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.gate_up_proj_weight_quantizers.203 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.gate_up_proj_weight_quantizers.204 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.gate_up_proj_weight_quantizers.205 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.gate_up_proj_weight_quantizers.206 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.gate_up_proj_weight_quantizers.207 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.17e-01 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.gate_up_proj_weight_quantizers.208 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.gate_up_proj_weight_quantizers.209 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.gate_up_proj_weight_quantizers.210 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.gate_up_proj_weight_quantizers.211 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.gate_up_proj_weight_quantizers.212 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.26e-01 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.gate_up_proj_weight_quantizers.213 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.gate_up_proj_weight_quantizers.214 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.82e-01 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.gate_up_proj_weight_quantizers.215 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.gate_up_proj_weight_quantizers.216 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.gate_up_proj_weight_quantizers.217 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.gate_up_proj_weight_quantizers.218 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.gate_up_proj_weight_quantizers.219 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.gate_up_proj_weight_quantizers.220 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.88e-01 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.gate_up_proj_weight_quantizers.221 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.gate_up_proj_weight_quantizers.222 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.gate_up_proj_weight_quantizers.223 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.gate_up_proj_weight_quantizers.224 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.gate_up_proj_weight_quantizers.225 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.gate_up_proj_weight_quantizers.226 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.gate_up_proj_weight_quantizers.227 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.gate_up_proj_weight_quantizers.228 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.85e-01 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.gate_up_proj_weight_quantizers.229 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.99e-01 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.gate_up_proj_weight_quantizers.230 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.gate_up_proj_weight_quantizers.231 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.58e-01 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.gate_up_proj_weight_quantizers.232 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.gate_up_proj_weight_quantizers.233 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.gate_up_proj_weight_quantizers.234 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.gate_up_proj_weight_quantizers.235 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.gate_up_proj_weight_quantizers.236 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.gate_up_proj_weight_quantizers.237 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.gate_up_proj_weight_quantizers.238 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.gate_up_proj_weight_quantizers.239 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.gate_up_proj_weight_quantizers.240 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.gate_up_proj_weight_quantizers.241 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.gate_up_proj_weight_quantizers.242 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.gate_up_proj_weight_quantizers.243 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.gate_up_proj_weight_quantizers.244 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.70e-01 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.gate_up_proj_weight_quantizers.245 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.gate_up_proj_weight_quantizers.246 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.62e-01 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.gate_up_proj_weight_quantizers.247 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.gate_up_proj_weight_quantizers.248 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.19e-01 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.gate_up_proj_weight_quantizers.249 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.gate_up_proj_weight_quantizers.250 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.gate_up_proj_weight_quantizers.251 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.gate_up_proj_weight_quantizers.252 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.gate_up_proj_weight_quantizers.253 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.gate_up_proj_weight_quantizers.254 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.73e-01 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.gate_up_proj_weight_quantizers.255 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.27e-01 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.down_proj_input_quantizer TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.44e+00 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.down_proj_weight_quantizers.0 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.64e-02 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.down_proj_weight_quantizers.1 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.down_proj_weight_quantizers.2 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.down_proj_weight_quantizers.3 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.72e-01 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.down_proj_weight_quantizers.4 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.down_proj_weight_quantizers.5 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.down_proj_weight_quantizers.6 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.22e-01 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.down_proj_weight_quantizers.7 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.down_proj_weight_quantizers.8 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.down_proj_weight_quantizers.9 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.down_proj_weight_quantizers.10 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.down_proj_weight_quantizers.11 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.down_proj_weight_quantizers.12 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.down_proj_weight_quantizers.13 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.down_proj_weight_quantizers.14 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.33e-01 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.down_proj_weight_quantizers.15 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.down_proj_weight_quantizers.16 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.28e-01 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.down_proj_weight_quantizers.17 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.down_proj_weight_quantizers.18 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.down_proj_weight_quantizers.19 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.down_proj_weight_quantizers.20 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.59e-02 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.down_proj_weight_quantizers.21 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.down_proj_weight_quantizers.22 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.45e-01 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.down_proj_weight_quantizers.23 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.down_proj_weight_quantizers.24 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.down_proj_weight_quantizers.25 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.53e-01 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.down_proj_weight_quantizers.26 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.down_proj_weight_quantizers.27 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.down_proj_weight_quantizers.28 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.down_proj_weight_quantizers.29 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.68e-01 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.down_proj_weight_quantizers.30 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.down_proj_weight_quantizers.31 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.28e-01 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.down_proj_weight_quantizers.32 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.down_proj_weight_quantizers.33 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.down_proj_weight_quantizers.34 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.down_proj_weight_quantizers.35 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.19e-01 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.down_proj_weight_quantizers.36 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.down_proj_weight_quantizers.37 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.down_proj_weight_quantizers.38 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.48e-01 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.down_proj_weight_quantizers.39 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.down_proj_weight_quantizers.40 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.down_proj_weight_quantizers.41 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.down_proj_weight_quantizers.42 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.down_proj_weight_quantizers.43 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.36e-01 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.down_proj_weight_quantizers.44 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.down_proj_weight_quantizers.45 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.down_proj_weight_quantizers.46 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.26e-01 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.down_proj_weight_quantizers.47 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.down_proj_weight_quantizers.48 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.20e-01 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.down_proj_weight_quantizers.49 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.down_proj_weight_quantizers.50 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.down_proj_weight_quantizers.51 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.down_proj_weight_quantizers.52 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.down_proj_weight_quantizers.53 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.down_proj_weight_quantizers.54 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.14e-01 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.down_proj_weight_quantizers.55 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.down_proj_weight_quantizers.56 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.down_proj_weight_quantizers.57 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.down_proj_weight_quantizers.58 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.down_proj_weight_quantizers.59 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.down_proj_weight_quantizers.60 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.82e-01 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.down_proj_weight_quantizers.61 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.20e-01 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.down_proj_weight_quantizers.62 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.down_proj_weight_quantizers.63 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.down_proj_weight_quantizers.64 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.26e-01 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.down_proj_weight_quantizers.65 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.down_proj_weight_quantizers.66 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.59e-02 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.down_proj_weight_quantizers.67 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.down_proj_weight_quantizers.68 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.down_proj_weight_quantizers.69 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.37e-01 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.down_proj_weight_quantizers.70 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.down_proj_weight_quantizers.71 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.down_proj_weight_quantizers.72 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.down_proj_weight_quantizers.73 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.down_proj_weight_quantizers.74 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.down_proj_weight_quantizers.75 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.down_proj_weight_quantizers.76 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.down_proj_weight_quantizers.77 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.down_proj_weight_quantizers.78 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.down_proj_weight_quantizers.79 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.down_proj_weight_quantizers.80 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.down_proj_weight_quantizers.81 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.down_proj_weight_quantizers.82 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.16e-01 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.down_proj_weight_quantizers.83 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.down_proj_weight_quantizers.84 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.down_proj_weight_quantizers.85 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.down_proj_weight_quantizers.86 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.27e-01 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.down_proj_weight_quantizers.87 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.48e-01 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.down_proj_weight_quantizers.88 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.down_proj_weight_quantizers.89 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.43e-01 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.down_proj_weight_quantizers.90 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.down_proj_weight_quantizers.91 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.15e-01 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.down_proj_weight_quantizers.92 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.down_proj_weight_quantizers.93 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.down_proj_weight_quantizers.94 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.down_proj_weight_quantizers.95 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.down_proj_weight_quantizers.96 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.down_proj_weight_quantizers.97 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.down_proj_weight_quantizers.98 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.down_proj_weight_quantizers.99 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.down_proj_weight_quantizers.100 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.15e-01 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.down_proj_weight_quantizers.101 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.down_proj_weight_quantizers.102 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.56e-01 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.down_proj_weight_quantizers.103 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.down_proj_weight_quantizers.104 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.down_proj_weight_quantizers.105 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.down_proj_weight_quantizers.106 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.down_proj_weight_quantizers.107 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.74e-01 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.down_proj_weight_quantizers.108 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.10e-01 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.down_proj_weight_quantizers.109 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.down_proj_weight_quantizers.110 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.down_proj_weight_quantizers.111 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.56e-01 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.down_proj_weight_quantizers.112 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.down_proj_weight_quantizers.113 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=4.47e-01 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.down_proj_weight_quantizers.114 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.down_proj_weight_quantizers.115 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.41e-01 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.down_proj_weight_quantizers.116 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.33e-01 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.down_proj_weight_quantizers.117 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.down_proj_weight_quantizers.118 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.down_proj_weight_quantizers.119 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.down_proj_weight_quantizers.120 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.down_proj_weight_quantizers.121 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.27e-01 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.down_proj_weight_quantizers.122 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.65e-01 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.down_proj_weight_quantizers.123 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.17e-01 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.down_proj_weight_quantizers.124 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.30e-01 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.down_proj_weight_quantizers.125 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.59e-01 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.down_proj_weight_quantizers.126 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.16e-01 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.down_proj_weight_quantizers.127 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.down_proj_weight_quantizers.128 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.down_proj_weight_quantizers.129 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.down_proj_weight_quantizers.130 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.44e-01 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.down_proj_weight_quantizers.131 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.down_proj_weight_quantizers.132 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.34e-01 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.down_proj_weight_quantizers.133 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.56e-01 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.down_proj_weight_quantizers.134 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.down_proj_weight_quantizers.135 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.down_proj_weight_quantizers.136 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.23e-01 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.down_proj_weight_quantizers.137 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.down_proj_weight_quantizers.138 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.down_proj_weight_quantizers.139 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.down_proj_weight_quantizers.140 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.down_proj_weight_quantizers.141 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.28e-01 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.down_proj_weight_quantizers.142 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.down_proj_weight_quantizers.143 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.down_proj_weight_quantizers.144 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.down_proj_weight_quantizers.145 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.down_proj_weight_quantizers.146 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.67e-01 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.down_proj_weight_quantizers.147 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.down_proj_weight_quantizers.148 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.28e-01 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.down_proj_weight_quantizers.149 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.down_proj_weight_quantizers.150 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.down_proj_weight_quantizers.151 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.down_proj_weight_quantizers.152 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.down_proj_weight_quantizers.153 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.30e-01 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.down_proj_weight_quantizers.154 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.35e-01 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.down_proj_weight_quantizers.155 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.16e-01 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.down_proj_weight_quantizers.156 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.down_proj_weight_quantizers.157 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.down_proj_weight_quantizers.158 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.down_proj_weight_quantizers.159 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.down_proj_weight_quantizers.160 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.down_proj_weight_quantizers.161 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.down_proj_weight_quantizers.162 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.down_proj_weight_quantizers.163 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.down_proj_weight_quantizers.164 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.42e-01 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.down_proj_weight_quantizers.165 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.down_proj_weight_quantizers.166 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.down_proj_weight_quantizers.167 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.down_proj_weight_quantizers.168 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.51e-01 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.down_proj_weight_quantizers.169 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.down_proj_weight_quantizers.170 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.down_proj_weight_quantizers.171 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.88e-01 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.down_proj_weight_quantizers.172 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.60e-01 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.down_proj_weight_quantizers.173 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.down_proj_weight_quantizers.174 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.down_proj_weight_quantizers.175 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.down_proj_weight_quantizers.176 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.down_proj_weight_quantizers.177 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.46e-01 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.down_proj_weight_quantizers.178 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.down_proj_weight_quantizers.179 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.down_proj_weight_quantizers.180 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.57e-01 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.down_proj_weight_quantizers.181 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.36e-01 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.down_proj_weight_quantizers.182 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.down_proj_weight_quantizers.183 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.51e-01 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.down_proj_weight_quantizers.184 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.43e-01 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.down_proj_weight_quantizers.185 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.down_proj_weight_quantizers.186 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.down_proj_weight_quantizers.187 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.68e-01 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.down_proj_weight_quantizers.188 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.down_proj_weight_quantizers.189 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.59e-02 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.down_proj_weight_quantizers.190 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.down_proj_weight_quantizers.191 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.down_proj_weight_quantizers.192 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.down_proj_weight_quantizers.193 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.59e-01 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.down_proj_weight_quantizers.194 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.41e-01 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.down_proj_weight_quantizers.195 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.17e-01 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.down_proj_weight_quantizers.196 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.25e-01 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.down_proj_weight_quantizers.197 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.32e-01 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.down_proj_weight_quantizers.198 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.down_proj_weight_quantizers.199 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.down_proj_weight_quantizers.200 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.25e-01 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.down_proj_weight_quantizers.201 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.19e-01 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.down_proj_weight_quantizers.202 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.down_proj_weight_quantizers.203 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.down_proj_weight_quantizers.204 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.down_proj_weight_quantizers.205 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.down_proj_weight_quantizers.206 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.down_proj_weight_quantizers.207 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.down_proj_weight_quantizers.208 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.down_proj_weight_quantizers.209 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.23e-01 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.down_proj_weight_quantizers.210 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.down_proj_weight_quantizers.211 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.14e-01 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.down_proj_weight_quantizers.212 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.62e-01 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.down_proj_weight_quantizers.213 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.down_proj_weight_quantizers.214 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.99e-01 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.down_proj_weight_quantizers.215 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.38e-01 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.down_proj_weight_quantizers.216 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.16e-01 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.down_proj_weight_quantizers.217 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.down_proj_weight_quantizers.218 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.49e-01 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.down_proj_weight_quantizers.219 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.55e-01 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.down_proj_weight_quantizers.220 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.down_proj_weight_quantizers.221 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.down_proj_weight_quantizers.222 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.down_proj_weight_quantizers.223 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.down_proj_weight_quantizers.224 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.down_proj_weight_quantizers.225 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.down_proj_weight_quantizers.226 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.down_proj_weight_quantizers.227 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.down_proj_weight_quantizers.228 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.52e-01 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.down_proj_weight_quantizers.229 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.21e-01 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.down_proj_weight_quantizers.230 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.down_proj_weight_quantizers.231 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.73e-01 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.down_proj_weight_quantizers.232 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.22e-01 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.down_proj_weight_quantizers.233 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.down_proj_weight_quantizers.234 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.down_proj_weight_quantizers.235 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.down_proj_weight_quantizers.236 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.down_proj_weight_quantizers.237 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.50e-02 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.down_proj_weight_quantizers.238 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.down_proj_weight_quantizers.239 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.down_proj_weight_quantizers.240 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.down_proj_weight_quantizers.241 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.21e-01 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.down_proj_weight_quantizers.242 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.down_proj_weight_quantizers.243 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.down_proj_weight_quantizers.244 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.down_proj_weight_quantizers.245 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.23e-01 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.down_proj_weight_quantizers.246 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.66e-01 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.down_proj_weight_quantizers.247 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.down_proj_weight_quantizers.248 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.15e-01 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.down_proj_weight_quantizers.249 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.50e-02 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.down_proj_weight_quantizers.250 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.18e-01 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.down_proj_weight_quantizers.251 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.down_proj_weight_quantizers.252 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.down_proj_weight_quantizers.253 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.down_proj_weight_quantizers.254 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.29e-01 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.down_proj_weight_quantizers.255 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.29e-01 calibrator=MaxCalibrator quant) +model.layers.16.mlp.shared_experts.gate_proj.input_quantizer TensorQuantizer(disabled) +model.layers.16.mlp.shared_experts.gate_proj.output_quantizer TensorQuantizer(disabled) +model.layers.16.mlp.shared_experts.gate_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.16.mlp.shared_experts.up_proj.input_quantizer TensorQuantizer(disabled) +model.layers.16.mlp.shared_experts.up_proj.output_quantizer TensorQuantizer(disabled) +model.layers.16.mlp.shared_experts.up_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.16.mlp.shared_experts.down_proj.input_quantizer TensorQuantizer(disabled) +model.layers.16.mlp.shared_experts.down_proj.output_quantizer TensorQuantizer(disabled) +model.layers.16.mlp.shared_experts.down_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.17.self_attn.q_a_proj.input_quantizer TensorQuantizer(disabled) +model.layers.17.self_attn.q_a_proj.output_quantizer TensorQuantizer(disabled) +model.layers.17.self_attn.q_a_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.17.self_attn.q_b_proj.input_quantizer TensorQuantizer(disabled) +model.layers.17.self_attn.q_b_proj.output_quantizer TensorQuantizer(disabled) +model.layers.17.self_attn.q_b_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.17.self_attn.kv_a_proj_with_mqa.input_quantizer TensorQuantizer(disabled) +model.layers.17.self_attn.kv_a_proj_with_mqa.output_quantizer TensorQuantizer(disabled) +model.layers.17.self_attn.kv_a_proj_with_mqa.weight_quantizer TensorQuantizer(disabled) +model.layers.17.self_attn.kv_b_proj.input_quantizer TensorQuantizer(disabled) +model.layers.17.self_attn.kv_b_proj.output_quantizer TensorQuantizer(disabled) +model.layers.17.self_attn.kv_b_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.17.self_attn.o_proj.input_quantizer TensorQuantizer(disabled) +model.layers.17.self_attn.o_proj.output_quantizer TensorQuantizer(disabled) +model.layers.17.self_attn.o_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.17.self_attn.q_bmm_quantizer TensorQuantizer(disabled) +model.layers.17.self_attn.k_bmm_quantizer TensorQuantizer((4, 3) bit fake per-tensor amax=4.48e+02(const) calibrator=MaxCalibrator quant) +model.layers.17.self_attn.v_bmm_quantizer TensorQuantizer((4, 3) bit fake per-tensor amax=4.48e+02(const) calibrator=MaxCalibrator quant) +model.layers.17.self_attn.p_bmm_quantizer TensorQuantizer(disabled) +model.layers.17.mlp.experts.gate_up_proj_input_quantizer TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=6.05e-01 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.gate_up_proj_weight_quantizers.0 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.gate_up_proj_weight_quantizers.1 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.gate_up_proj_weight_quantizers.2 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.gate_up_proj_weight_quantizers.3 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.gate_up_proj_weight_quantizers.4 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.gate_up_proj_weight_quantizers.5 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.gate_up_proj_weight_quantizers.6 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.gate_up_proj_weight_quantizers.7 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.gate_up_proj_weight_quantizers.8 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.gate_up_proj_weight_quantizers.9 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.87e-01 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.gate_up_proj_weight_quantizers.10 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.gate_up_proj_weight_quantizers.11 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.gate_up_proj_weight_quantizers.12 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.30e-01 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.gate_up_proj_weight_quantizers.13 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.gate_up_proj_weight_quantizers.14 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.gate_up_proj_weight_quantizers.15 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.gate_up_proj_weight_quantizers.16 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.gate_up_proj_weight_quantizers.17 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.gate_up_proj_weight_quantizers.18 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.70e-01 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.gate_up_proj_weight_quantizers.19 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.61e-01 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.gate_up_proj_weight_quantizers.20 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.gate_up_proj_weight_quantizers.21 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.gate_up_proj_weight_quantizers.22 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.24e-01 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.gate_up_proj_weight_quantizers.23 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.gate_up_proj_weight_quantizers.24 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.gate_up_proj_weight_quantizers.25 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.gate_up_proj_weight_quantizers.26 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.gate_up_proj_weight_quantizers.27 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.gate_up_proj_weight_quantizers.28 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.gate_up_proj_weight_quantizers.29 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.gate_up_proj_weight_quantizers.30 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.gate_up_proj_weight_quantizers.31 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.gate_up_proj_weight_quantizers.32 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.40e-01 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.gate_up_proj_weight_quantizers.33 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.gate_up_proj_weight_quantizers.34 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.gate_up_proj_weight_quantizers.35 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.gate_up_proj_weight_quantizers.36 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.gate_up_proj_weight_quantizers.37 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.gate_up_proj_weight_quantizers.38 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.gate_up_proj_weight_quantizers.39 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.gate_up_proj_weight_quantizers.40 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.71e-01 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.gate_up_proj_weight_quantizers.41 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.gate_up_proj_weight_quantizers.42 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.17e-01 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.gate_up_proj_weight_quantizers.43 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.gate_up_proj_weight_quantizers.44 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.44e-01 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.gate_up_proj_weight_quantizers.45 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.gate_up_proj_weight_quantizers.46 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.gate_up_proj_weight_quantizers.47 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.gate_up_proj_weight_quantizers.48 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.gate_up_proj_weight_quantizers.49 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.gate_up_proj_weight_quantizers.50 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.03e-01 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.gate_up_proj_weight_quantizers.51 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.gate_up_proj_weight_quantizers.52 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.gate_up_proj_weight_quantizers.53 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.gate_up_proj_weight_quantizers.54 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.gate_up_proj_weight_quantizers.55 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.32e-01 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.gate_up_proj_weight_quantizers.56 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.gate_up_proj_weight_quantizers.57 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.gate_up_proj_weight_quantizers.58 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.gate_up_proj_weight_quantizers.59 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.gate_up_proj_weight_quantizers.60 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.gate_up_proj_weight_quantizers.61 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.gate_up_proj_weight_quantizers.62 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.gate_up_proj_weight_quantizers.63 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.gate_up_proj_weight_quantizers.64 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.gate_up_proj_weight_quantizers.65 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.gate_up_proj_weight_quantizers.66 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.gate_up_proj_weight_quantizers.67 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.45e-02 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.gate_up_proj_weight_quantizers.68 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.gate_up_proj_weight_quantizers.69 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.gate_up_proj_weight_quantizers.70 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.gate_up_proj_weight_quantizers.71 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.30e-01 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.gate_up_proj_weight_quantizers.72 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.gate_up_proj_weight_quantizers.73 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.21e-01 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.gate_up_proj_weight_quantizers.74 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.gate_up_proj_weight_quantizers.75 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.gate_up_proj_weight_quantizers.76 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.75e-01 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.gate_up_proj_weight_quantizers.77 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.14e-01 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.gate_up_proj_weight_quantizers.78 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.gate_up_proj_weight_quantizers.79 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.gate_up_proj_weight_quantizers.80 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.gate_up_proj_weight_quantizers.81 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.gate_up_proj_weight_quantizers.82 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.gate_up_proj_weight_quantizers.83 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.gate_up_proj_weight_quantizers.84 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.gate_up_proj_weight_quantizers.85 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.gate_up_proj_weight_quantizers.86 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.gate_up_proj_weight_quantizers.87 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.gate_up_proj_weight_quantizers.88 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.gate_up_proj_weight_quantizers.89 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.36e-01 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.gate_up_proj_weight_quantizers.90 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.gate_up_proj_weight_quantizers.91 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.gate_up_proj_weight_quantizers.92 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.gate_up_proj_weight_quantizers.93 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.gate_up_proj_weight_quantizers.94 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.gate_up_proj_weight_quantizers.95 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.gate_up_proj_weight_quantizers.96 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.gate_up_proj_weight_quantizers.97 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.gate_up_proj_weight_quantizers.98 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.gate_up_proj_weight_quantizers.99 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.gate_up_proj_weight_quantizers.100 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.gate_up_proj_weight_quantizers.101 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.gate_up_proj_weight_quantizers.102 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.gate_up_proj_weight_quantizers.103 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.gate_up_proj_weight_quantizers.104 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.gate_up_proj_weight_quantizers.105 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.gate_up_proj_weight_quantizers.106 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.gate_up_proj_weight_quantizers.107 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.gate_up_proj_weight_quantizers.108 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.gate_up_proj_weight_quantizers.109 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.gate_up_proj_weight_quantizers.110 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.gate_up_proj_weight_quantizers.111 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.gate_up_proj_weight_quantizers.112 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.gate_up_proj_weight_quantizers.113 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.gate_up_proj_weight_quantizers.114 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.18e-01 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.gate_up_proj_weight_quantizers.115 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.26e-01 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.gate_up_proj_weight_quantizers.116 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.gate_up_proj_weight_quantizers.117 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.gate_up_proj_weight_quantizers.118 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.gate_up_proj_weight_quantizers.119 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.gate_up_proj_weight_quantizers.120 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.19e-01 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.gate_up_proj_weight_quantizers.121 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.13e-01 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.gate_up_proj_weight_quantizers.122 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.gate_up_proj_weight_quantizers.123 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.gate_up_proj_weight_quantizers.124 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.gate_up_proj_weight_quantizers.125 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.gate_up_proj_weight_quantizers.126 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.gate_up_proj_weight_quantizers.127 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.gate_up_proj_weight_quantizers.128 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.35e-01 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.gate_up_proj_weight_quantizers.129 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.08e-01 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.gate_up_proj_weight_quantizers.130 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.16e-01 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.gate_up_proj_weight_quantizers.131 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.gate_up_proj_weight_quantizers.132 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.gate_up_proj_weight_quantizers.133 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.gate_up_proj_weight_quantizers.134 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.gate_up_proj_weight_quantizers.135 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.51e-01 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.gate_up_proj_weight_quantizers.136 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.gate_up_proj_weight_quantizers.137 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.gate_up_proj_weight_quantizers.138 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.gate_up_proj_weight_quantizers.139 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.gate_up_proj_weight_quantizers.140 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.gate_up_proj_weight_quantizers.141 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.gate_up_proj_weight_quantizers.142 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.gate_up_proj_weight_quantizers.143 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.gate_up_proj_weight_quantizers.144 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.gate_up_proj_weight_quantizers.145 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.gate_up_proj_weight_quantizers.146 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.gate_up_proj_weight_quantizers.147 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.gate_up_proj_weight_quantizers.148 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.gate_up_proj_weight_quantizers.149 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.gate_up_proj_weight_quantizers.150 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.gate_up_proj_weight_quantizers.151 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.gate_up_proj_weight_quantizers.152 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.21e-01 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.gate_up_proj_weight_quantizers.153 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.89e-01 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.gate_up_proj_weight_quantizers.154 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.gate_up_proj_weight_quantizers.155 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.gate_up_proj_weight_quantizers.156 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.gate_up_proj_weight_quantizers.157 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.20e-01 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.gate_up_proj_weight_quantizers.158 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.gate_up_proj_weight_quantizers.159 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.gate_up_proj_weight_quantizers.160 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.gate_up_proj_weight_quantizers.161 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.gate_up_proj_weight_quantizers.162 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.gate_up_proj_weight_quantizers.163 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.gate_up_proj_weight_quantizers.164 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.gate_up_proj_weight_quantizers.165 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.gate_up_proj_weight_quantizers.166 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.gate_up_proj_weight_quantizers.167 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.gate_up_proj_weight_quantizers.168 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.74e-01 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.gate_up_proj_weight_quantizers.169 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.13e-01 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.gate_up_proj_weight_quantizers.170 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.gate_up_proj_weight_quantizers.171 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.gate_up_proj_weight_quantizers.172 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.02e-01 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.gate_up_proj_weight_quantizers.173 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.15e-01 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.gate_up_proj_weight_quantizers.174 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.gate_up_proj_weight_quantizers.175 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.gate_up_proj_weight_quantizers.176 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.gate_up_proj_weight_quantizers.177 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.gate_up_proj_weight_quantizers.178 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.gate_up_proj_weight_quantizers.179 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.gate_up_proj_weight_quantizers.180 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.46e-01 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.gate_up_proj_weight_quantizers.181 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.gate_up_proj_weight_quantizers.182 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.gate_up_proj_weight_quantizers.183 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.32e-01 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.gate_up_proj_weight_quantizers.184 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.gate_up_proj_weight_quantizers.185 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.gate_up_proj_weight_quantizers.186 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.gate_up_proj_weight_quantizers.187 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.gate_up_proj_weight_quantizers.188 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.46e-01 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.gate_up_proj_weight_quantizers.189 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.gate_up_proj_weight_quantizers.190 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.67e-01 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.gate_up_proj_weight_quantizers.191 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.gate_up_proj_weight_quantizers.192 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.gate_up_proj_weight_quantizers.193 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.gate_up_proj_weight_quantizers.194 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.gate_up_proj_weight_quantizers.195 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.gate_up_proj_weight_quantizers.196 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.gate_up_proj_weight_quantizers.197 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.gate_up_proj_weight_quantizers.198 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.gate_up_proj_weight_quantizers.199 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.gate_up_proj_weight_quantizers.200 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.gate_up_proj_weight_quantizers.201 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.gate_up_proj_weight_quantizers.202 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.gate_up_proj_weight_quantizers.203 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.gate_up_proj_weight_quantizers.204 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.gate_up_proj_weight_quantizers.205 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.22e-01 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.gate_up_proj_weight_quantizers.206 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.gate_up_proj_weight_quantizers.207 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.gate_up_proj_weight_quantizers.208 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.31e-01 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.gate_up_proj_weight_quantizers.209 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.28e-01 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.gate_up_proj_weight_quantizers.210 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.gate_up_proj_weight_quantizers.211 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.gate_up_proj_weight_quantizers.212 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.gate_up_proj_weight_quantizers.213 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.gate_up_proj_weight_quantizers.214 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.28e-01 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.gate_up_proj_weight_quantizers.215 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.28e-01 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.gate_up_proj_weight_quantizers.216 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.gate_up_proj_weight_quantizers.217 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.gate_up_proj_weight_quantizers.218 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.46e-01 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.gate_up_proj_weight_quantizers.219 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.gate_up_proj_weight_quantizers.220 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.gate_up_proj_weight_quantizers.221 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.gate_up_proj_weight_quantizers.222 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.gate_up_proj_weight_quantizers.223 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.30e-01 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.gate_up_proj_weight_quantizers.224 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.40e-01 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.gate_up_proj_weight_quantizers.225 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.gate_up_proj_weight_quantizers.226 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.gate_up_proj_weight_quantizers.227 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.gate_up_proj_weight_quantizers.228 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.gate_up_proj_weight_quantizers.229 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.gate_up_proj_weight_quantizers.230 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.62e-01 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.gate_up_proj_weight_quantizers.231 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.gate_up_proj_weight_quantizers.232 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.gate_up_proj_weight_quantizers.233 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.22e-01 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.gate_up_proj_weight_quantizers.234 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.gate_up_proj_weight_quantizers.235 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.gate_up_proj_weight_quantizers.236 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.gate_up_proj_weight_quantizers.237 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.gate_up_proj_weight_quantizers.238 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.72e-01 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.gate_up_proj_weight_quantizers.239 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.gate_up_proj_weight_quantizers.240 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.70e-01 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.gate_up_proj_weight_quantizers.241 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.gate_up_proj_weight_quantizers.242 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.gate_up_proj_weight_quantizers.243 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.gate_up_proj_weight_quantizers.244 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.gate_up_proj_weight_quantizers.245 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.gate_up_proj_weight_quantizers.246 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.23e-01 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.gate_up_proj_weight_quantizers.247 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.gate_up_proj_weight_quantizers.248 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.gate_up_proj_weight_quantizers.249 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.gate_up_proj_weight_quantizers.250 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.06e-01 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.gate_up_proj_weight_quantizers.251 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.14e-01 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.gate_up_proj_weight_quantizers.252 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.gate_up_proj_weight_quantizers.253 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.gate_up_proj_weight_quantizers.254 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.32e-01 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.gate_up_proj_weight_quantizers.255 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.down_proj_input_quantizer TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.41e+00 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.down_proj_weight_quantizers.0 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.down_proj_weight_quantizers.1 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.31e-01 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.down_proj_weight_quantizers.2 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.down_proj_weight_quantizers.3 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.down_proj_weight_quantizers.4 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.down_proj_weight_quantizers.5 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.down_proj_weight_quantizers.6 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.down_proj_weight_quantizers.7 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.19e-01 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.down_proj_weight_quantizers.8 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.down_proj_weight_quantizers.9 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.17e-01 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.down_proj_weight_quantizers.10 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.22e-01 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.down_proj_weight_quantizers.11 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.down_proj_weight_quantizers.12 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.68e-01 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.down_proj_weight_quantizers.13 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.24e-01 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.down_proj_weight_quantizers.14 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.down_proj_weight_quantizers.15 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.down_proj_weight_quantizers.16 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.down_proj_weight_quantizers.17 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.16e-01 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.down_proj_weight_quantizers.18 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.58e-01 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.down_proj_weight_quantizers.19 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.56e-01 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.down_proj_weight_quantizers.20 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.down_proj_weight_quantizers.21 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.70e-01 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.down_proj_weight_quantizers.22 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.down_proj_weight_quantizers.23 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.down_proj_weight_quantizers.24 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.down_proj_weight_quantizers.25 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.down_proj_weight_quantizers.26 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.28e-01 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.down_proj_weight_quantizers.27 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.down_proj_weight_quantizers.28 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.down_proj_weight_quantizers.29 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.down_proj_weight_quantizers.30 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.down_proj_weight_quantizers.31 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.down_proj_weight_quantizers.32 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.44e-01 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.down_proj_weight_quantizers.33 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.down_proj_weight_quantizers.34 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.30e-01 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.down_proj_weight_quantizers.35 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.down_proj_weight_quantizers.36 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.down_proj_weight_quantizers.37 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.down_proj_weight_quantizers.38 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.down_proj_weight_quantizers.39 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.down_proj_weight_quantizers.40 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.down_proj_weight_quantizers.41 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.down_proj_weight_quantizers.42 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.down_proj_weight_quantizers.43 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.down_proj_weight_quantizers.44 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.25e-01 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.down_proj_weight_quantizers.45 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.down_proj_weight_quantizers.46 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.down_proj_weight_quantizers.47 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.down_proj_weight_quantizers.48 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.down_proj_weight_quantizers.49 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.down_proj_weight_quantizers.50 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.81e-01 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.down_proj_weight_quantizers.51 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.16e-01 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.down_proj_weight_quantizers.52 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.down_proj_weight_quantizers.53 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.down_proj_weight_quantizers.54 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.down_proj_weight_quantizers.55 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.69e-01 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.down_proj_weight_quantizers.56 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.30e-01 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.down_proj_weight_quantizers.57 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.down_proj_weight_quantizers.58 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.down_proj_weight_quantizers.59 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.down_proj_weight_quantizers.60 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.down_proj_weight_quantizers.61 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.down_proj_weight_quantizers.62 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.down_proj_weight_quantizers.63 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.41e-01 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.down_proj_weight_quantizers.64 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.down_proj_weight_quantizers.65 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.down_proj_weight_quantizers.66 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.down_proj_weight_quantizers.67 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.45e-02 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.down_proj_weight_quantizers.68 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.down_proj_weight_quantizers.69 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.down_proj_weight_quantizers.70 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.45e-01 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.down_proj_weight_quantizers.71 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.51e-01 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.down_proj_weight_quantizers.72 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.down_proj_weight_quantizers.73 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.down_proj_weight_quantizers.74 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.down_proj_weight_quantizers.75 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.down_proj_weight_quantizers.76 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.30e-01 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.down_proj_weight_quantizers.77 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.down_proj_weight_quantizers.78 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.down_proj_weight_quantizers.79 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.down_proj_weight_quantizers.80 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.down_proj_weight_quantizers.81 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.down_proj_weight_quantizers.82 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.down_proj_weight_quantizers.83 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.down_proj_weight_quantizers.84 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.down_proj_weight_quantizers.85 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.down_proj_weight_quantizers.86 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.down_proj_weight_quantizers.87 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.84e-01 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.down_proj_weight_quantizers.88 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.down_proj_weight_quantizers.89 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.52e-01 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.down_proj_weight_quantizers.90 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.down_proj_weight_quantizers.91 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.down_proj_weight_quantizers.92 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.down_proj_weight_quantizers.93 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.down_proj_weight_quantizers.94 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.down_proj_weight_quantizers.95 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.down_proj_weight_quantizers.96 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.down_proj_weight_quantizers.97 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.down_proj_weight_quantizers.98 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.down_proj_weight_quantizers.99 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.down_proj_weight_quantizers.100 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.down_proj_weight_quantizers.101 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.down_proj_weight_quantizers.102 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.54e-02 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.down_proj_weight_quantizers.103 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.17e-01 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.down_proj_weight_quantizers.104 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.down_proj_weight_quantizers.105 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.down_proj_weight_quantizers.106 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.down_proj_weight_quantizers.107 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.down_proj_weight_quantizers.108 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.73e-01 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.down_proj_weight_quantizers.109 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.down_proj_weight_quantizers.110 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.down_proj_weight_quantizers.111 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.down_proj_weight_quantizers.112 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.down_proj_weight_quantizers.113 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.down_proj_weight_quantizers.114 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.84e-01 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.down_proj_weight_quantizers.115 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.90e-01 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.down_proj_weight_quantizers.116 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.down_proj_weight_quantizers.117 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.down_proj_weight_quantizers.118 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.down_proj_weight_quantizers.119 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.down_proj_weight_quantizers.120 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.down_proj_weight_quantizers.121 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.down_proj_weight_quantizers.122 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.14e-01 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.down_proj_weight_quantizers.123 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.down_proj_weight_quantizers.124 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.15e-01 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.down_proj_weight_quantizers.125 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.down_proj_weight_quantizers.126 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.down_proj_weight_quantizers.127 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.down_proj_weight_quantizers.128 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.33e-01 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.down_proj_weight_quantizers.129 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.down_proj_weight_quantizers.130 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.down_proj_weight_quantizers.131 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.down_proj_weight_quantizers.132 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.down_proj_weight_quantizers.133 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.down_proj_weight_quantizers.134 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.down_proj_weight_quantizers.135 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.41e-01 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.down_proj_weight_quantizers.136 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.33e-01 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.down_proj_weight_quantizers.137 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.89e-01 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.down_proj_weight_quantizers.138 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.down_proj_weight_quantizers.139 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.down_proj_weight_quantizers.140 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.down_proj_weight_quantizers.141 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.down_proj_weight_quantizers.142 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.down_proj_weight_quantizers.143 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.down_proj_weight_quantizers.144 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.down_proj_weight_quantizers.145 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.down_proj_weight_quantizers.146 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.down_proj_weight_quantizers.147 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.54e-02 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.down_proj_weight_quantizers.148 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.down_proj_weight_quantizers.149 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.down_proj_weight_quantizers.150 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.down_proj_weight_quantizers.151 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.down_proj_weight_quantizers.152 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.43e-01 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.down_proj_weight_quantizers.153 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.43e-01 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.down_proj_weight_quantizers.154 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.down_proj_weight_quantizers.155 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.down_proj_weight_quantizers.156 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.down_proj_weight_quantizers.157 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.down_proj_weight_quantizers.158 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.down_proj_weight_quantizers.159 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.down_proj_weight_quantizers.160 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.down_proj_weight_quantizers.161 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.53e-01 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.down_proj_weight_quantizers.162 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.down_proj_weight_quantizers.163 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.31e-01 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.down_proj_weight_quantizers.164 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.down_proj_weight_quantizers.165 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.down_proj_weight_quantizers.166 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.down_proj_weight_quantizers.167 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.down_proj_weight_quantizers.168 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.27e-01 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.down_proj_weight_quantizers.169 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.65e-01 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.down_proj_weight_quantizers.170 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.down_proj_weight_quantizers.171 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.down_proj_weight_quantizers.172 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.66e-01 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.down_proj_weight_quantizers.173 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.30e-01 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.down_proj_weight_quantizers.174 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.40e-02 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.down_proj_weight_quantizers.175 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.24e-01 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.down_proj_weight_quantizers.176 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.71e-01 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.down_proj_weight_quantizers.177 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.18e-01 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.down_proj_weight_quantizers.178 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.down_proj_weight_quantizers.179 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.down_proj_weight_quantizers.180 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.23e-01 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.down_proj_weight_quantizers.181 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.down_proj_weight_quantizers.182 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.down_proj_weight_quantizers.183 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.down_proj_weight_quantizers.184 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.72e-01 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.down_proj_weight_quantizers.185 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.down_proj_weight_quantizers.186 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.down_proj_weight_quantizers.187 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.30e-01 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.down_proj_weight_quantizers.188 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.16e-01 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.down_proj_weight_quantizers.189 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.down_proj_weight_quantizers.190 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.90e-01 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.down_proj_weight_quantizers.191 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.down_proj_weight_quantizers.192 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.down_proj_weight_quantizers.193 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.down_proj_weight_quantizers.194 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.down_proj_weight_quantizers.195 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.down_proj_weight_quantizers.196 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.down_proj_weight_quantizers.197 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.50e-02 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.down_proj_weight_quantizers.198 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.down_proj_weight_quantizers.199 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.down_proj_weight_quantizers.200 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.down_proj_weight_quantizers.201 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.down_proj_weight_quantizers.202 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.down_proj_weight_quantizers.203 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.down_proj_weight_quantizers.204 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.59e-02 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.down_proj_weight_quantizers.205 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.down_proj_weight_quantizers.206 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.down_proj_weight_quantizers.207 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.21e-01 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.down_proj_weight_quantizers.208 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.down_proj_weight_quantizers.209 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.35e-01 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.down_proj_weight_quantizers.210 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.down_proj_weight_quantizers.211 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.down_proj_weight_quantizers.212 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.down_proj_weight_quantizers.213 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.down_proj_weight_quantizers.214 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.92e-01 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.down_proj_weight_quantizers.215 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.29e-01 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.down_proj_weight_quantizers.216 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.down_proj_weight_quantizers.217 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.17e-01 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.down_proj_weight_quantizers.218 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.down_proj_weight_quantizers.219 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.40e-01 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.down_proj_weight_quantizers.220 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.down_proj_weight_quantizers.221 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.down_proj_weight_quantizers.222 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.down_proj_weight_quantizers.223 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.25e-01 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.down_proj_weight_quantizers.224 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.39e-01 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.down_proj_weight_quantizers.225 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.down_proj_weight_quantizers.226 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.64e-01 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.down_proj_weight_quantizers.227 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.down_proj_weight_quantizers.228 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.down_proj_weight_quantizers.229 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.down_proj_weight_quantizers.230 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.35e-01 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.down_proj_weight_quantizers.231 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.14e-01 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.down_proj_weight_quantizers.232 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.down_proj_weight_quantizers.233 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.down_proj_weight_quantizers.234 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.down_proj_weight_quantizers.235 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.down_proj_weight_quantizers.236 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.down_proj_weight_quantizers.237 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.down_proj_weight_quantizers.238 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.27e-01 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.down_proj_weight_quantizers.239 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.down_proj_weight_quantizers.240 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.down_proj_weight_quantizers.241 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.down_proj_weight_quantizers.242 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.down_proj_weight_quantizers.243 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.down_proj_weight_quantizers.244 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.down_proj_weight_quantizers.245 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.39e-01 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.down_proj_weight_quantizers.246 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.50e-01 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.down_proj_weight_quantizers.247 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.down_proj_weight_quantizers.248 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.20e-01 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.down_proj_weight_quantizers.249 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.down_proj_weight_quantizers.250 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.68e-01 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.down_proj_weight_quantizers.251 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.33e-01 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.down_proj_weight_quantizers.252 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.down_proj_weight_quantizers.253 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.41e-01 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.down_proj_weight_quantizers.254 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.36e-01 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.down_proj_weight_quantizers.255 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.17.mlp.shared_experts.gate_proj.input_quantizer TensorQuantizer(disabled) +model.layers.17.mlp.shared_experts.gate_proj.output_quantizer TensorQuantizer(disabled) +model.layers.17.mlp.shared_experts.gate_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.17.mlp.shared_experts.up_proj.input_quantizer TensorQuantizer(disabled) +model.layers.17.mlp.shared_experts.up_proj.output_quantizer TensorQuantizer(disabled) +model.layers.17.mlp.shared_experts.up_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.17.mlp.shared_experts.down_proj.input_quantizer TensorQuantizer(disabled) +model.layers.17.mlp.shared_experts.down_proj.output_quantizer TensorQuantizer(disabled) +model.layers.17.mlp.shared_experts.down_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.18.self_attn.q_a_proj.input_quantizer TensorQuantizer(disabled) +model.layers.18.self_attn.q_a_proj.output_quantizer TensorQuantizer(disabled) +model.layers.18.self_attn.q_a_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.18.self_attn.q_b_proj.input_quantizer TensorQuantizer(disabled) +model.layers.18.self_attn.q_b_proj.output_quantizer TensorQuantizer(disabled) +model.layers.18.self_attn.q_b_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.18.self_attn.kv_a_proj_with_mqa.input_quantizer TensorQuantizer(disabled) +model.layers.18.self_attn.kv_a_proj_with_mqa.output_quantizer TensorQuantizer(disabled) +model.layers.18.self_attn.kv_a_proj_with_mqa.weight_quantizer TensorQuantizer(disabled) +model.layers.18.self_attn.kv_b_proj.input_quantizer TensorQuantizer(disabled) +model.layers.18.self_attn.kv_b_proj.output_quantizer TensorQuantizer(disabled) +model.layers.18.self_attn.kv_b_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.18.self_attn.o_proj.input_quantizer TensorQuantizer(disabled) +model.layers.18.self_attn.o_proj.output_quantizer TensorQuantizer(disabled) +model.layers.18.self_attn.o_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.18.self_attn.indexer.wq_b.input_quantizer TensorQuantizer(disabled) +model.layers.18.self_attn.indexer.wq_b.output_quantizer TensorQuantizer(disabled) +model.layers.18.self_attn.indexer.wq_b.weight_quantizer TensorQuantizer(disabled) +model.layers.18.self_attn.indexer.wk.input_quantizer TensorQuantizer(disabled) +model.layers.18.self_attn.indexer.wk.output_quantizer TensorQuantizer(disabled) +model.layers.18.self_attn.indexer.wk.weight_quantizer TensorQuantizer(disabled) +model.layers.18.self_attn.indexer.k_norm.input_quantizer TensorQuantizer(disabled) +model.layers.18.self_attn.indexer.k_norm.output_quantizer TensorQuantizer(disabled) +model.layers.18.self_attn.indexer.weights_proj.input_quantizer TensorQuantizer(disabled) +model.layers.18.self_attn.indexer.weights_proj.output_quantizer TensorQuantizer(disabled) +model.layers.18.self_attn.indexer.weights_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.18.self_attn.q_bmm_quantizer TensorQuantizer(disabled) +model.layers.18.self_attn.k_bmm_quantizer TensorQuantizer((4, 3) bit fake per-tensor amax=4.48e+02(const) calibrator=MaxCalibrator quant) +model.layers.18.self_attn.v_bmm_quantizer TensorQuantizer((4, 3) bit fake per-tensor amax=4.48e+02(const) calibrator=MaxCalibrator quant) +model.layers.18.self_attn.p_bmm_quantizer TensorQuantizer(disabled) +model.layers.18.mlp.experts.gate_up_proj_input_quantizer TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=6.29e-01 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.gate_up_proj_weight_quantizers.0 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.gate_up_proj_weight_quantizers.1 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.gate_up_proj_weight_quantizers.2 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.gate_up_proj_weight_quantizers.3 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.gate_up_proj_weight_quantizers.4 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.gate_up_proj_weight_quantizers.5 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.gate_up_proj_weight_quantizers.6 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.gate_up_proj_weight_quantizers.7 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.gate_up_proj_weight_quantizers.8 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.gate_up_proj_weight_quantizers.9 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.gate_up_proj_weight_quantizers.10 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.gate_up_proj_weight_quantizers.11 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.gate_up_proj_weight_quantizers.12 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.gate_up_proj_weight_quantizers.13 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.gate_up_proj_weight_quantizers.14 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.gate_up_proj_weight_quantizers.15 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.gate_up_proj_weight_quantizers.16 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.gate_up_proj_weight_quantizers.17 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.gate_up_proj_weight_quantizers.18 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.gate_up_proj_weight_quantizers.19 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.gate_up_proj_weight_quantizers.20 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.gate_up_proj_weight_quantizers.21 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.gate_up_proj_weight_quantizers.22 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.gate_up_proj_weight_quantizers.23 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.99e-01 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.gate_up_proj_weight_quantizers.24 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.gate_up_proj_weight_quantizers.25 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.gate_up_proj_weight_quantizers.26 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.gate_up_proj_weight_quantizers.27 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.gate_up_proj_weight_quantizers.28 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.gate_up_proj_weight_quantizers.29 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.gate_up_proj_weight_quantizers.30 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.gate_up_proj_weight_quantizers.31 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.gate_up_proj_weight_quantizers.32 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.gate_up_proj_weight_quantizers.33 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.37e-01 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.gate_up_proj_weight_quantizers.34 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.gate_up_proj_weight_quantizers.35 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.gate_up_proj_weight_quantizers.36 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.gate_up_proj_weight_quantizers.37 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.gate_up_proj_weight_quantizers.38 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.gate_up_proj_weight_quantizers.39 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.gate_up_proj_weight_quantizers.40 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.36e-01 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.gate_up_proj_weight_quantizers.41 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.gate_up_proj_weight_quantizers.42 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.gate_up_proj_weight_quantizers.43 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.gate_up_proj_weight_quantizers.44 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.gate_up_proj_weight_quantizers.45 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.gate_up_proj_weight_quantizers.46 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.gate_up_proj_weight_quantizers.47 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.25e-01 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.gate_up_proj_weight_quantizers.48 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.gate_up_proj_weight_quantizers.49 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.gate_up_proj_weight_quantizers.50 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.gate_up_proj_weight_quantizers.51 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.gate_up_proj_weight_quantizers.52 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.gate_up_proj_weight_quantizers.53 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.gate_up_proj_weight_quantizers.54 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.gate_up_proj_weight_quantizers.55 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.gate_up_proj_weight_quantizers.56 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.gate_up_proj_weight_quantizers.57 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.gate_up_proj_weight_quantizers.58 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.gate_up_proj_weight_quantizers.59 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.gate_up_proj_weight_quantizers.60 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.gate_up_proj_weight_quantizers.61 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.29e-01 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.gate_up_proj_weight_quantizers.62 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.gate_up_proj_weight_quantizers.63 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.gate_up_proj_weight_quantizers.64 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.gate_up_proj_weight_quantizers.65 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.gate_up_proj_weight_quantizers.66 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.gate_up_proj_weight_quantizers.67 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.gate_up_proj_weight_quantizers.68 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.gate_up_proj_weight_quantizers.69 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.gate_up_proj_weight_quantizers.70 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.gate_up_proj_weight_quantizers.71 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.gate_up_proj_weight_quantizers.72 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.gate_up_proj_weight_quantizers.73 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.gate_up_proj_weight_quantizers.74 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.gate_up_proj_weight_quantizers.75 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.17e-01 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.gate_up_proj_weight_quantizers.76 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.gate_up_proj_weight_quantizers.77 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.gate_up_proj_weight_quantizers.78 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.gate_up_proj_weight_quantizers.79 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.gate_up_proj_weight_quantizers.80 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.gate_up_proj_weight_quantizers.81 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.gate_up_proj_weight_quantizers.82 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.gate_up_proj_weight_quantizers.83 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.gate_up_proj_weight_quantizers.84 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.gate_up_proj_weight_quantizers.85 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.gate_up_proj_weight_quantizers.86 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.gate_up_proj_weight_quantizers.87 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.gate_up_proj_weight_quantizers.88 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.62e-01 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.gate_up_proj_weight_quantizers.89 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.64e-01 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.gate_up_proj_weight_quantizers.90 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.gate_up_proj_weight_quantizers.91 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.gate_up_proj_weight_quantizers.92 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.gate_up_proj_weight_quantizers.93 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.gate_up_proj_weight_quantizers.94 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.gate_up_proj_weight_quantizers.95 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.gate_up_proj_weight_quantizers.96 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.gate_up_proj_weight_quantizers.97 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.gate_up_proj_weight_quantizers.98 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.gate_up_proj_weight_quantizers.99 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.gate_up_proj_weight_quantizers.100 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.gate_up_proj_weight_quantizers.101 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.gate_up_proj_weight_quantizers.102 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.gate_up_proj_weight_quantizers.103 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.gate_up_proj_weight_quantizers.104 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.gate_up_proj_weight_quantizers.105 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.13e-01 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.gate_up_proj_weight_quantizers.106 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.gate_up_proj_weight_quantizers.107 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.gate_up_proj_weight_quantizers.108 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.gate_up_proj_weight_quantizers.109 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.25e-01 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.gate_up_proj_weight_quantizers.110 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.gate_up_proj_weight_quantizers.111 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.gate_up_proj_weight_quantizers.112 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.gate_up_proj_weight_quantizers.113 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.gate_up_proj_weight_quantizers.114 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.gate_up_proj_weight_quantizers.115 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.gate_up_proj_weight_quantizers.116 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.gate_up_proj_weight_quantizers.117 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.14e-01 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.gate_up_proj_weight_quantizers.118 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.gate_up_proj_weight_quantizers.119 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.gate_up_proj_weight_quantizers.120 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.gate_up_proj_weight_quantizers.121 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.gate_up_proj_weight_quantizers.122 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.gate_up_proj_weight_quantizers.123 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.gate_up_proj_weight_quantizers.124 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.gate_up_proj_weight_quantizers.125 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.gate_up_proj_weight_quantizers.126 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.gate_up_proj_weight_quantizers.127 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.gate_up_proj_weight_quantizers.128 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.gate_up_proj_weight_quantizers.129 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.gate_up_proj_weight_quantizers.130 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.gate_up_proj_weight_quantizers.131 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.gate_up_proj_weight_quantizers.132 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.gate_up_proj_weight_quantizers.133 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.gate_up_proj_weight_quantizers.134 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.gate_up_proj_weight_quantizers.135 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.gate_up_proj_weight_quantizers.136 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.gate_up_proj_weight_quantizers.137 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.gate_up_proj_weight_quantizers.138 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.gate_up_proj_weight_quantizers.139 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.gate_up_proj_weight_quantizers.140 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.gate_up_proj_weight_quantizers.141 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.gate_up_proj_weight_quantizers.142 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.gate_up_proj_weight_quantizers.143 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.gate_up_proj_weight_quantizers.144 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.gate_up_proj_weight_quantizers.145 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.gate_up_proj_weight_quantizers.146 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.gate_up_proj_weight_quantizers.147 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.gate_up_proj_weight_quantizers.148 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.gate_up_proj_weight_quantizers.149 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.gate_up_proj_weight_quantizers.150 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.gate_up_proj_weight_quantizers.151 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.gate_up_proj_weight_quantizers.152 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.gate_up_proj_weight_quantizers.153 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.gate_up_proj_weight_quantizers.154 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.gate_up_proj_weight_quantizers.155 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.gate_up_proj_weight_quantizers.156 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.gate_up_proj_weight_quantizers.157 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.gate_up_proj_weight_quantizers.158 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.gate_up_proj_weight_quantizers.159 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.gate_up_proj_weight_quantizers.160 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.gate_up_proj_weight_quantizers.161 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.29e-01 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.gate_up_proj_weight_quantizers.162 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.gate_up_proj_weight_quantizers.163 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.gate_up_proj_weight_quantizers.164 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.gate_up_proj_weight_quantizers.165 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.gate_up_proj_weight_quantizers.166 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.gate_up_proj_weight_quantizers.167 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.gate_up_proj_weight_quantizers.168 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.gate_up_proj_weight_quantizers.169 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.gate_up_proj_weight_quantizers.170 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.gate_up_proj_weight_quantizers.171 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.gate_up_proj_weight_quantizers.172 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.gate_up_proj_weight_quantizers.173 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.gate_up_proj_weight_quantizers.174 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.gate_up_proj_weight_quantizers.175 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.gate_up_proj_weight_quantizers.176 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.gate_up_proj_weight_quantizers.177 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.gate_up_proj_weight_quantizers.178 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.gate_up_proj_weight_quantizers.179 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.gate_up_proj_weight_quantizers.180 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.gate_up_proj_weight_quantizers.181 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.gate_up_proj_weight_quantizers.182 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.gate_up_proj_weight_quantizers.183 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.gate_up_proj_weight_quantizers.184 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.gate_up_proj_weight_quantizers.185 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.18e-01 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.gate_up_proj_weight_quantizers.186 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.gate_up_proj_weight_quantizers.187 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.gate_up_proj_weight_quantizers.188 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.gate_up_proj_weight_quantizers.189 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.20e-01 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.gate_up_proj_weight_quantizers.190 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.gate_up_proj_weight_quantizers.191 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.gate_up_proj_weight_quantizers.192 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.86e-01 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.gate_up_proj_weight_quantizers.193 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.gate_up_proj_weight_quantizers.194 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.gate_up_proj_weight_quantizers.195 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.gate_up_proj_weight_quantizers.196 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.14e-01 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.gate_up_proj_weight_quantizers.197 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.gate_up_proj_weight_quantizers.198 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.gate_up_proj_weight_quantizers.199 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.59e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.gate_up_proj_weight_quantizers.200 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.gate_up_proj_weight_quantizers.201 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.gate_up_proj_weight_quantizers.202 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.gate_up_proj_weight_quantizers.203 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.gate_up_proj_weight_quantizers.204 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.gate_up_proj_weight_quantizers.205 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.17e-01 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.gate_up_proj_weight_quantizers.206 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.gate_up_proj_weight_quantizers.207 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.gate_up_proj_weight_quantizers.208 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.gate_up_proj_weight_quantizers.209 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.gate_up_proj_weight_quantizers.210 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.gate_up_proj_weight_quantizers.211 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.gate_up_proj_weight_quantizers.212 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.gate_up_proj_weight_quantizers.213 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.gate_up_proj_weight_quantizers.214 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.gate_up_proj_weight_quantizers.215 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.gate_up_proj_weight_quantizers.216 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.gate_up_proj_weight_quantizers.217 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.gate_up_proj_weight_quantizers.218 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.gate_up_proj_weight_quantizers.219 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.gate_up_proj_weight_quantizers.220 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.gate_up_proj_weight_quantizers.221 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.gate_up_proj_weight_quantizers.222 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.gate_up_proj_weight_quantizers.223 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.gate_up_proj_weight_quantizers.224 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.gate_up_proj_weight_quantizers.225 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.gate_up_proj_weight_quantizers.226 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.gate_up_proj_weight_quantizers.227 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.gate_up_proj_weight_quantizers.228 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.gate_up_proj_weight_quantizers.229 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.gate_up_proj_weight_quantizers.230 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.gate_up_proj_weight_quantizers.231 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.gate_up_proj_weight_quantizers.232 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.23e-01 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.gate_up_proj_weight_quantizers.233 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.25e-01 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.gate_up_proj_weight_quantizers.234 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.gate_up_proj_weight_quantizers.235 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.gate_up_proj_weight_quantizers.236 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.gate_up_proj_weight_quantizers.237 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.gate_up_proj_weight_quantizers.238 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.gate_up_proj_weight_quantizers.239 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.gate_up_proj_weight_quantizers.240 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.16e-01 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.gate_up_proj_weight_quantizers.241 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.gate_up_proj_weight_quantizers.242 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.gate_up_proj_weight_quantizers.243 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.gate_up_proj_weight_quantizers.244 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.gate_up_proj_weight_quantizers.245 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.gate_up_proj_weight_quantizers.246 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.gate_up_proj_weight_quantizers.247 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.gate_up_proj_weight_quantizers.248 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.gate_up_proj_weight_quantizers.249 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.gate_up_proj_weight_quantizers.250 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.gate_up_proj_weight_quantizers.251 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.gate_up_proj_weight_quantizers.252 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.gate_up_proj_weight_quantizers.253 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.gate_up_proj_weight_quantizers.254 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.gate_up_proj_weight_quantizers.255 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.down_proj_input_quantizer TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=4.62e+00 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.down_proj_weight_quantizers.0 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.down_proj_weight_quantizers.1 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.down_proj_weight_quantizers.2 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.down_proj_weight_quantizers.3 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.down_proj_weight_quantizers.4 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.54e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.down_proj_weight_quantizers.5 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.64e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.down_proj_weight_quantizers.6 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.59e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.down_proj_weight_quantizers.7 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.down_proj_weight_quantizers.8 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.down_proj_weight_quantizers.9 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.down_proj_weight_quantizers.10 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.down_proj_weight_quantizers.11 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.13e-01 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.down_proj_weight_quantizers.12 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.down_proj_weight_quantizers.13 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.down_proj_weight_quantizers.14 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.down_proj_weight_quantizers.15 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.down_proj_weight_quantizers.16 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.down_proj_weight_quantizers.17 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.50e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.down_proj_weight_quantizers.18 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.down_proj_weight_quantizers.19 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.down_proj_weight_quantizers.20 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.down_proj_weight_quantizers.21 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.down_proj_weight_quantizers.22 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.down_proj_weight_quantizers.23 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.down_proj_weight_quantizers.24 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.down_proj_weight_quantizers.25 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.down_proj_weight_quantizers.26 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.down_proj_weight_quantizers.27 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.down_proj_weight_quantizers.28 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.79e-01 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.down_proj_weight_quantizers.29 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.down_proj_weight_quantizers.30 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.down_proj_weight_quantizers.31 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.down_proj_weight_quantizers.32 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.down_proj_weight_quantizers.33 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.10e-01 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.down_proj_weight_quantizers.34 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.down_proj_weight_quantizers.35 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.down_proj_weight_quantizers.36 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.40e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.down_proj_weight_quantizers.37 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.down_proj_weight_quantizers.38 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.43e-01 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.down_proj_weight_quantizers.39 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.down_proj_weight_quantizers.40 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.down_proj_weight_quantizers.41 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.down_proj_weight_quantizers.42 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.down_proj_weight_quantizers.43 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.down_proj_weight_quantizers.44 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.20e-01 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.down_proj_weight_quantizers.45 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.down_proj_weight_quantizers.46 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.down_proj_weight_quantizers.47 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.down_proj_weight_quantizers.48 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.down_proj_weight_quantizers.49 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.down_proj_weight_quantizers.50 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.down_proj_weight_quantizers.51 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.down_proj_weight_quantizers.52 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.down_proj_weight_quantizers.53 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.down_proj_weight_quantizers.54 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.down_proj_weight_quantizers.55 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.down_proj_weight_quantizers.56 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.down_proj_weight_quantizers.57 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.down_proj_weight_quantizers.58 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.down_proj_weight_quantizers.59 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.down_proj_weight_quantizers.60 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.45e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.down_proj_weight_quantizers.61 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.36e-01 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.down_proj_weight_quantizers.62 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.64e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.down_proj_weight_quantizers.63 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.down_proj_weight_quantizers.64 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.down_proj_weight_quantizers.65 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.down_proj_weight_quantizers.66 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.down_proj_weight_quantizers.67 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.down_proj_weight_quantizers.68 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.40e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.down_proj_weight_quantizers.69 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.30e-01 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.down_proj_weight_quantizers.70 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.down_proj_weight_quantizers.71 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.down_proj_weight_quantizers.72 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.down_proj_weight_quantizers.73 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.36e-01 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.down_proj_weight_quantizers.74 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.down_proj_weight_quantizers.75 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.26e-01 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.down_proj_weight_quantizers.76 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.down_proj_weight_quantizers.77 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.down_proj_weight_quantizers.78 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.60e-01 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.down_proj_weight_quantizers.79 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.80e-01 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.down_proj_weight_quantizers.80 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.64e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.down_proj_weight_quantizers.81 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.down_proj_weight_quantizers.82 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.64e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.down_proj_weight_quantizers.83 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.down_proj_weight_quantizers.84 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.down_proj_weight_quantizers.85 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.down_proj_weight_quantizers.86 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.59e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.down_proj_weight_quantizers.87 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.down_proj_weight_quantizers.88 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.49e-01 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.down_proj_weight_quantizers.89 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.19e-01 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.down_proj_weight_quantizers.90 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.64e-01 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.down_proj_weight_quantizers.91 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.98e-01 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.down_proj_weight_quantizers.92 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.down_proj_weight_quantizers.93 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.down_proj_weight_quantizers.94 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.down_proj_weight_quantizers.95 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.50e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.down_proj_weight_quantizers.96 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.down_proj_weight_quantizers.97 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.45e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.down_proj_weight_quantizers.98 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.down_proj_weight_quantizers.99 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.down_proj_weight_quantizers.100 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.down_proj_weight_quantizers.101 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.down_proj_weight_quantizers.102 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.54e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.down_proj_weight_quantizers.103 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.down_proj_weight_quantizers.104 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.down_proj_weight_quantizers.105 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.34e-01 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.down_proj_weight_quantizers.106 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.down_proj_weight_quantizers.107 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.down_proj_weight_quantizers.108 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.13e-01 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.down_proj_weight_quantizers.109 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.down_proj_weight_quantizers.110 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.down_proj_weight_quantizers.111 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.down_proj_weight_quantizers.112 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.down_proj_weight_quantizers.113 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.down_proj_weight_quantizers.114 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.down_proj_weight_quantizers.115 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.down_proj_weight_quantizers.116 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.down_proj_weight_quantizers.117 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.down_proj_weight_quantizers.118 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.down_proj_weight_quantizers.119 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.50e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.down_proj_weight_quantizers.120 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.down_proj_weight_quantizers.121 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.down_proj_weight_quantizers.122 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.down_proj_weight_quantizers.123 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.down_proj_weight_quantizers.124 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.down_proj_weight_quantizers.125 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.64e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.down_proj_weight_quantizers.126 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.down_proj_weight_quantizers.127 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.down_proj_weight_quantizers.128 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.down_proj_weight_quantizers.129 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.down_proj_weight_quantizers.130 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.down_proj_weight_quantizers.131 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.down_proj_weight_quantizers.132 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.down_proj_weight_quantizers.133 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.down_proj_weight_quantizers.134 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.down_proj_weight_quantizers.135 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.31e-01 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.down_proj_weight_quantizers.136 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.down_proj_weight_quantizers.137 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.down_proj_weight_quantizers.138 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.down_proj_weight_quantizers.139 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.down_proj_weight_quantizers.140 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.13e-01 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.down_proj_weight_quantizers.141 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.down_proj_weight_quantizers.142 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.down_proj_weight_quantizers.143 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.41e-01 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.down_proj_weight_quantizers.144 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.64e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.down_proj_weight_quantizers.145 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.down_proj_weight_quantizers.146 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.down_proj_weight_quantizers.147 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.down_proj_weight_quantizers.148 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.down_proj_weight_quantizers.149 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.64e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.down_proj_weight_quantizers.150 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.down_proj_weight_quantizers.151 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.down_proj_weight_quantizers.152 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.down_proj_weight_quantizers.153 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.14e-01 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.down_proj_weight_quantizers.154 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.down_proj_weight_quantizers.155 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.down_proj_weight_quantizers.156 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.down_proj_weight_quantizers.157 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.down_proj_weight_quantizers.158 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.down_proj_weight_quantizers.159 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.down_proj_weight_quantizers.160 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.down_proj_weight_quantizers.161 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.46e-01 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.down_proj_weight_quantizers.162 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.down_proj_weight_quantizers.163 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.59e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.down_proj_weight_quantizers.164 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.down_proj_weight_quantizers.165 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.59e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.down_proj_weight_quantizers.166 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.down_proj_weight_quantizers.167 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.down_proj_weight_quantizers.168 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.down_proj_weight_quantizers.169 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.down_proj_weight_quantizers.170 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.64e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.down_proj_weight_quantizers.171 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.down_proj_weight_quantizers.172 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.down_proj_weight_quantizers.173 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.down_proj_weight_quantizers.174 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.down_proj_weight_quantizers.175 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.down_proj_weight_quantizers.176 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.down_proj_weight_quantizers.177 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.down_proj_weight_quantizers.178 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.down_proj_weight_quantizers.179 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.down_proj_weight_quantizers.180 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.down_proj_weight_quantizers.181 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.down_proj_weight_quantizers.182 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.down_proj_weight_quantizers.183 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.down_proj_weight_quantizers.184 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.50e-01 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.down_proj_weight_quantizers.185 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.62e-01 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.down_proj_weight_quantizers.186 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.down_proj_weight_quantizers.187 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.down_proj_weight_quantizers.188 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.down_proj_weight_quantizers.189 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.down_proj_weight_quantizers.190 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.25e-01 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.down_proj_weight_quantizers.191 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.down_proj_weight_quantizers.192 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.down_proj_weight_quantizers.193 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.down_proj_weight_quantizers.194 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.down_proj_weight_quantizers.195 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.down_proj_weight_quantizers.196 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.down_proj_weight_quantizers.197 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.40e-01 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.down_proj_weight_quantizers.198 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.54e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.down_proj_weight_quantizers.199 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.down_proj_weight_quantizers.200 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.down_proj_weight_quantizers.201 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.down_proj_weight_quantizers.202 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.down_proj_weight_quantizers.203 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.down_proj_weight_quantizers.204 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.down_proj_weight_quantizers.205 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.26e-01 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.down_proj_weight_quantizers.206 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.down_proj_weight_quantizers.207 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.64e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.down_proj_weight_quantizers.208 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.down_proj_weight_quantizers.209 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.64e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.down_proj_weight_quantizers.210 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.down_proj_weight_quantizers.211 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.down_proj_weight_quantizers.212 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.down_proj_weight_quantizers.213 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.59e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.down_proj_weight_quantizers.214 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.38e-01 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.down_proj_weight_quantizers.215 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.down_proj_weight_quantizers.216 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.down_proj_weight_quantizers.217 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.down_proj_weight_quantizers.218 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.down_proj_weight_quantizers.219 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.down_proj_weight_quantizers.220 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.down_proj_weight_quantizers.221 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.down_proj_weight_quantizers.222 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.down_proj_weight_quantizers.223 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.down_proj_weight_quantizers.224 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.down_proj_weight_quantizers.225 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.down_proj_weight_quantizers.226 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.down_proj_weight_quantizers.227 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.down_proj_weight_quantizers.228 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.50e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.down_proj_weight_quantizers.229 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.down_proj_weight_quantizers.230 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.down_proj_weight_quantizers.231 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.down_proj_weight_quantizers.232 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.down_proj_weight_quantizers.233 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.54e-01 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.down_proj_weight_quantizers.234 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.54e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.down_proj_weight_quantizers.235 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.23e-01 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.down_proj_weight_quantizers.236 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.down_proj_weight_quantizers.237 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.down_proj_weight_quantizers.238 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.down_proj_weight_quantizers.239 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.down_proj_weight_quantizers.240 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.down_proj_weight_quantizers.241 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.down_proj_weight_quantizers.242 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.down_proj_weight_quantizers.243 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.down_proj_weight_quantizers.244 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.down_proj_weight_quantizers.245 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.down_proj_weight_quantizers.246 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.down_proj_weight_quantizers.247 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.down_proj_weight_quantizers.248 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.down_proj_weight_quantizers.249 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.54e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.down_proj_weight_quantizers.250 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.64e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.down_proj_weight_quantizers.251 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.down_proj_weight_quantizers.252 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.down_proj_weight_quantizers.253 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.down_proj_weight_quantizers.254 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.down_proj_weight_quantizers.255 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.shared_experts.gate_proj.input_quantizer TensorQuantizer(disabled) +model.layers.18.mlp.shared_experts.gate_proj.output_quantizer TensorQuantizer(disabled) +model.layers.18.mlp.shared_experts.gate_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.18.mlp.shared_experts.up_proj.input_quantizer TensorQuantizer(disabled) +model.layers.18.mlp.shared_experts.up_proj.output_quantizer TensorQuantizer(disabled) +model.layers.18.mlp.shared_experts.up_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.18.mlp.shared_experts.down_proj.input_quantizer TensorQuantizer(disabled) +model.layers.18.mlp.shared_experts.down_proj.output_quantizer TensorQuantizer(disabled) +model.layers.18.mlp.shared_experts.down_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.19.self_attn.q_a_proj.input_quantizer TensorQuantizer(disabled) +model.layers.19.self_attn.q_a_proj.output_quantizer TensorQuantizer(disabled) +model.layers.19.self_attn.q_a_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.19.self_attn.q_b_proj.input_quantizer TensorQuantizer(disabled) +model.layers.19.self_attn.q_b_proj.output_quantizer TensorQuantizer(disabled) +model.layers.19.self_attn.q_b_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.19.self_attn.kv_a_proj_with_mqa.input_quantizer TensorQuantizer(disabled) +model.layers.19.self_attn.kv_a_proj_with_mqa.output_quantizer TensorQuantizer(disabled) +model.layers.19.self_attn.kv_a_proj_with_mqa.weight_quantizer TensorQuantizer(disabled) +model.layers.19.self_attn.kv_b_proj.input_quantizer TensorQuantizer(disabled) +model.layers.19.self_attn.kv_b_proj.output_quantizer TensorQuantizer(disabled) +model.layers.19.self_attn.kv_b_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.19.self_attn.o_proj.input_quantizer TensorQuantizer(disabled) +model.layers.19.self_attn.o_proj.output_quantizer TensorQuantizer(disabled) +model.layers.19.self_attn.o_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.19.self_attn.q_bmm_quantizer TensorQuantizer(disabled) +model.layers.19.self_attn.k_bmm_quantizer TensorQuantizer((4, 3) bit fake per-tensor amax=4.48e+02(const) calibrator=MaxCalibrator quant) +model.layers.19.self_attn.v_bmm_quantizer TensorQuantizer((4, 3) bit fake per-tensor amax=4.48e+02(const) calibrator=MaxCalibrator quant) +model.layers.19.self_attn.p_bmm_quantizer TensorQuantizer(disabled) +model.layers.19.mlp.experts.gate_up_proj_input_quantizer TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=6.33e-01 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.gate_up_proj_weight_quantizers.0 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.gate_up_proj_weight_quantizers.1 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.gate_up_proj_weight_quantizers.2 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.gate_up_proj_weight_quantizers.3 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.32e-01 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.gate_up_proj_weight_quantizers.4 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.89e-01 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.gate_up_proj_weight_quantizers.5 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.gate_up_proj_weight_quantizers.6 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.gate_up_proj_weight_quantizers.7 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.52e-01 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.gate_up_proj_weight_quantizers.8 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.gate_up_proj_weight_quantizers.9 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.gate_up_proj_weight_quantizers.10 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.gate_up_proj_weight_quantizers.11 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.gate_up_proj_weight_quantizers.12 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.gate_up_proj_weight_quantizers.13 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.gate_up_proj_weight_quantizers.14 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.29e-01 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.gate_up_proj_weight_quantizers.15 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.gate_up_proj_weight_quantizers.16 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.gate_up_proj_weight_quantizers.17 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.gate_up_proj_weight_quantizers.18 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.gate_up_proj_weight_quantizers.19 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.25e-01 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.gate_up_proj_weight_quantizers.20 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.gate_up_proj_weight_quantizers.21 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.gate_up_proj_weight_quantizers.22 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.gate_up_proj_weight_quantizers.23 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.gate_up_proj_weight_quantizers.24 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.gate_up_proj_weight_quantizers.25 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.gate_up_proj_weight_quantizers.26 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.gate_up_proj_weight_quantizers.27 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.gate_up_proj_weight_quantizers.28 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.26e-01 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.gate_up_proj_weight_quantizers.29 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.gate_up_proj_weight_quantizers.30 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.gate_up_proj_weight_quantizers.31 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.gate_up_proj_weight_quantizers.32 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.gate_up_proj_weight_quantizers.33 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.gate_up_proj_weight_quantizers.34 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.61e-01 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.gate_up_proj_weight_quantizers.35 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.85e-01 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.gate_up_proj_weight_quantizers.36 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.gate_up_proj_weight_quantizers.37 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.gate_up_proj_weight_quantizers.38 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.gate_up_proj_weight_quantizers.39 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.gate_up_proj_weight_quantizers.40 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.64e-01 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.gate_up_proj_weight_quantizers.41 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.gate_up_proj_weight_quantizers.42 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.gate_up_proj_weight_quantizers.43 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.99e-01 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.gate_up_proj_weight_quantizers.44 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.gate_up_proj_weight_quantizers.45 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.gate_up_proj_weight_quantizers.46 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.gate_up_proj_weight_quantizers.47 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.gate_up_proj_weight_quantizers.48 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.52e-01 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.gate_up_proj_weight_quantizers.49 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.gate_up_proj_weight_quantizers.50 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.gate_up_proj_weight_quantizers.51 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.gate_up_proj_weight_quantizers.52 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.31e-01 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.gate_up_proj_weight_quantizers.53 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.gate_up_proj_weight_quantizers.54 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.gate_up_proj_weight_quantizers.55 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.gate_up_proj_weight_quantizers.56 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.gate_up_proj_weight_quantizers.57 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.gate_up_proj_weight_quantizers.58 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.gate_up_proj_weight_quantizers.59 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.gate_up_proj_weight_quantizers.60 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.gate_up_proj_weight_quantizers.61 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.gate_up_proj_weight_quantizers.62 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.gate_up_proj_weight_quantizers.63 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.45e-01 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.gate_up_proj_weight_quantizers.64 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.gate_up_proj_weight_quantizers.65 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.gate_up_proj_weight_quantizers.66 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.gate_up_proj_weight_quantizers.67 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.20e-01 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.gate_up_proj_weight_quantizers.68 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.gate_up_proj_weight_quantizers.69 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.gate_up_proj_weight_quantizers.70 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.gate_up_proj_weight_quantizers.71 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.gate_up_proj_weight_quantizers.72 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.22e-01 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.gate_up_proj_weight_quantizers.73 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.gate_up_proj_weight_quantizers.74 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.gate_up_proj_weight_quantizers.75 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.gate_up_proj_weight_quantizers.76 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.gate_up_proj_weight_quantizers.77 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.gate_up_proj_weight_quantizers.78 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=4.10e-01 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.gate_up_proj_weight_quantizers.79 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.gate_up_proj_weight_quantizers.80 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.gate_up_proj_weight_quantizers.81 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.gate_up_proj_weight_quantizers.82 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.gate_up_proj_weight_quantizers.83 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.gate_up_proj_weight_quantizers.84 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.gate_up_proj_weight_quantizers.85 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.68e-01 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.gate_up_proj_weight_quantizers.86 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.gate_up_proj_weight_quantizers.87 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.04e-01 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.gate_up_proj_weight_quantizers.88 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.gate_up_proj_weight_quantizers.89 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.gate_up_proj_weight_quantizers.90 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.gate_up_proj_weight_quantizers.91 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.gate_up_proj_weight_quantizers.92 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.gate_up_proj_weight_quantizers.93 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.10e-01 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.gate_up_proj_weight_quantizers.94 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.32e-01 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.gate_up_proj_weight_quantizers.95 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.gate_up_proj_weight_quantizers.96 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.gate_up_proj_weight_quantizers.97 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.gate_up_proj_weight_quantizers.98 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.gate_up_proj_weight_quantizers.99 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.gate_up_proj_weight_quantizers.100 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.gate_up_proj_weight_quantizers.101 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.gate_up_proj_weight_quantizers.102 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.gate_up_proj_weight_quantizers.103 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.gate_up_proj_weight_quantizers.104 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.gate_up_proj_weight_quantizers.105 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.gate_up_proj_weight_quantizers.106 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.gate_up_proj_weight_quantizers.107 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.gate_up_proj_weight_quantizers.108 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.gate_up_proj_weight_quantizers.109 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.gate_up_proj_weight_quantizers.110 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.18e-01 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.gate_up_proj_weight_quantizers.111 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.gate_up_proj_weight_quantizers.112 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.gate_up_proj_weight_quantizers.113 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.gate_up_proj_weight_quantizers.114 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.gate_up_proj_weight_quantizers.115 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.gate_up_proj_weight_quantizers.116 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.15e-01 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.gate_up_proj_weight_quantizers.117 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.gate_up_proj_weight_quantizers.118 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.18e-01 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.gate_up_proj_weight_quantizers.119 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.gate_up_proj_weight_quantizers.120 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.gate_up_proj_weight_quantizers.121 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.gate_up_proj_weight_quantizers.122 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.gate_up_proj_weight_quantizers.123 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.gate_up_proj_weight_quantizers.124 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.gate_up_proj_weight_quantizers.125 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.34e-01 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.gate_up_proj_weight_quantizers.126 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.39e-01 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.gate_up_proj_weight_quantizers.127 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.gate_up_proj_weight_quantizers.128 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.gate_up_proj_weight_quantizers.129 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.gate_up_proj_weight_quantizers.130 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.gate_up_proj_weight_quantizers.131 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.gate_up_proj_weight_quantizers.132 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.gate_up_proj_weight_quantizers.133 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.gate_up_proj_weight_quantizers.134 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.gate_up_proj_weight_quantizers.135 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.gate_up_proj_weight_quantizers.136 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.gate_up_proj_weight_quantizers.137 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.23e-01 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.gate_up_proj_weight_quantizers.138 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.23e-01 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.gate_up_proj_weight_quantizers.139 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.gate_up_proj_weight_quantizers.140 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.gate_up_proj_weight_quantizers.141 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.gate_up_proj_weight_quantizers.142 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.gate_up_proj_weight_quantizers.143 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.56e-01 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.gate_up_proj_weight_quantizers.144 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.gate_up_proj_weight_quantizers.145 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.gate_up_proj_weight_quantizers.146 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.gate_up_proj_weight_quantizers.147 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.gate_up_proj_weight_quantizers.148 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.gate_up_proj_weight_quantizers.149 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.gate_up_proj_weight_quantizers.150 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.gate_up_proj_weight_quantizers.151 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.gate_up_proj_weight_quantizers.152 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.gate_up_proj_weight_quantizers.153 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.gate_up_proj_weight_quantizers.154 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.gate_up_proj_weight_quantizers.155 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.gate_up_proj_weight_quantizers.156 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.19e-01 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.gate_up_proj_weight_quantizers.157 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.gate_up_proj_weight_quantizers.158 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.gate_up_proj_weight_quantizers.159 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.gate_up_proj_weight_quantizers.160 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.gate_up_proj_weight_quantizers.161 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.gate_up_proj_weight_quantizers.162 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.gate_up_proj_weight_quantizers.163 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.59e-01 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.gate_up_proj_weight_quantizers.164 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.09e-01 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.gate_up_proj_weight_quantizers.165 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.gate_up_proj_weight_quantizers.166 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.gate_up_proj_weight_quantizers.167 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.gate_up_proj_weight_quantizers.168 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.gate_up_proj_weight_quantizers.169 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.gate_up_proj_weight_quantizers.170 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.gate_up_proj_weight_quantizers.171 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.gate_up_proj_weight_quantizers.172 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.23e-01 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.gate_up_proj_weight_quantizers.173 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.gate_up_proj_weight_quantizers.174 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.26e-01 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.gate_up_proj_weight_quantizers.175 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.19e-01 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.gate_up_proj_weight_quantizers.176 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.gate_up_proj_weight_quantizers.177 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.gate_up_proj_weight_quantizers.178 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.gate_up_proj_weight_quantizers.179 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.gate_up_proj_weight_quantizers.180 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.gate_up_proj_weight_quantizers.181 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.gate_up_proj_weight_quantizers.182 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.34e-01 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.gate_up_proj_weight_quantizers.183 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.gate_up_proj_weight_quantizers.184 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.gate_up_proj_weight_quantizers.185 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.gate_up_proj_weight_quantizers.186 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.70e-01 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.gate_up_proj_weight_quantizers.187 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.gate_up_proj_weight_quantizers.188 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.gate_up_proj_weight_quantizers.189 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.gate_up_proj_weight_quantizers.190 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.gate_up_proj_weight_quantizers.191 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.gate_up_proj_weight_quantizers.192 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.gate_up_proj_weight_quantizers.193 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.gate_up_proj_weight_quantizers.194 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.gate_up_proj_weight_quantizers.195 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.gate_up_proj_weight_quantizers.196 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.47e-01 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.gate_up_proj_weight_quantizers.197 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.69e-01 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.gate_up_proj_weight_quantizers.198 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.13e-01 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.gate_up_proj_weight_quantizers.199 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.gate_up_proj_weight_quantizers.200 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.gate_up_proj_weight_quantizers.201 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.gate_up_proj_weight_quantizers.202 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.gate_up_proj_weight_quantizers.203 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.gate_up_proj_weight_quantizers.204 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.gate_up_proj_weight_quantizers.205 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.28e-01 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.gate_up_proj_weight_quantizers.206 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.gate_up_proj_weight_quantizers.207 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.49e-01 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.gate_up_proj_weight_quantizers.208 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.gate_up_proj_weight_quantizers.209 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.30e-01 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.gate_up_proj_weight_quantizers.210 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.gate_up_proj_weight_quantizers.211 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.gate_up_proj_weight_quantizers.212 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.30e-01 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.gate_up_proj_weight_quantizers.213 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.43e-01 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.gate_up_proj_weight_quantizers.214 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.gate_up_proj_weight_quantizers.215 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.99e-01 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.gate_up_proj_weight_quantizers.216 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.13e-01 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.gate_up_proj_weight_quantizers.217 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.gate_up_proj_weight_quantizers.218 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.gate_up_proj_weight_quantizers.219 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.gate_up_proj_weight_quantizers.220 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.gate_up_proj_weight_quantizers.221 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.gate_up_proj_weight_quantizers.222 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.gate_up_proj_weight_quantizers.223 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.14e-01 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.gate_up_proj_weight_quantizers.224 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.gate_up_proj_weight_quantizers.225 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.49e-01 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.gate_up_proj_weight_quantizers.226 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.gate_up_proj_weight_quantizers.227 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.gate_up_proj_weight_quantizers.228 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.28e-01 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.gate_up_proj_weight_quantizers.229 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.gate_up_proj_weight_quantizers.230 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.gate_up_proj_weight_quantizers.231 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.gate_up_proj_weight_quantizers.232 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.39e-01 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.gate_up_proj_weight_quantizers.233 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.gate_up_proj_weight_quantizers.234 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.gate_up_proj_weight_quantizers.235 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.gate_up_proj_weight_quantizers.236 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.gate_up_proj_weight_quantizers.237 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.gate_up_proj_weight_quantizers.238 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.40e-01 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.gate_up_proj_weight_quantizers.239 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.39e-01 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.gate_up_proj_weight_quantizers.240 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.gate_up_proj_weight_quantizers.241 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.gate_up_proj_weight_quantizers.242 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.gate_up_proj_weight_quantizers.243 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.gate_up_proj_weight_quantizers.244 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.gate_up_proj_weight_quantizers.245 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.gate_up_proj_weight_quantizers.246 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.gate_up_proj_weight_quantizers.247 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.gate_up_proj_weight_quantizers.248 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.gate_up_proj_weight_quantizers.249 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.gate_up_proj_weight_quantizers.250 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.gate_up_proj_weight_quantizers.251 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.gate_up_proj_weight_quantizers.252 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.gate_up_proj_weight_quantizers.253 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.23e-01 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.gate_up_proj_weight_quantizers.254 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.gate_up_proj_weight_quantizers.255 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.73e-01 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.down_proj_input_quantizer TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.69e+00 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.down_proj_weight_quantizers.0 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.down_proj_weight_quantizers.1 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.down_proj_weight_quantizers.2 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.down_proj_weight_quantizers.3 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.44e-01 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.down_proj_weight_quantizers.4 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.down_proj_weight_quantizers.5 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.down_proj_weight_quantizers.6 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.50e-02 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.down_proj_weight_quantizers.7 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.down_proj_weight_quantizers.8 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.down_proj_weight_quantizers.9 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.18e-01 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.down_proj_weight_quantizers.10 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.down_proj_weight_quantizers.11 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.down_proj_weight_quantizers.12 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.down_proj_weight_quantizers.13 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.down_proj_weight_quantizers.14 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.59e-02 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.down_proj_weight_quantizers.15 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.64e-02 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.down_proj_weight_quantizers.16 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.down_proj_weight_quantizers.17 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.down_proj_weight_quantizers.18 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.down_proj_weight_quantizers.19 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.down_proj_weight_quantizers.20 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.down_proj_weight_quantizers.21 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.down_proj_weight_quantizers.22 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.64e-02 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.down_proj_weight_quantizers.23 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.down_proj_weight_quantizers.24 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.21e-01 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.down_proj_weight_quantizers.25 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.54e-02 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.down_proj_weight_quantizers.26 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.down_proj_weight_quantizers.27 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.26e-01 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.down_proj_weight_quantizers.28 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.39e-01 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.down_proj_weight_quantizers.29 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.down_proj_weight_quantizers.30 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.down_proj_weight_quantizers.31 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.down_proj_weight_quantizers.32 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.down_proj_weight_quantizers.33 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.25e-02 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.down_proj_weight_quantizers.34 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.54e-01 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.down_proj_weight_quantizers.35 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.31e-01 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.down_proj_weight_quantizers.36 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.down_proj_weight_quantizers.37 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.down_proj_weight_quantizers.38 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.down_proj_weight_quantizers.39 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.down_proj_weight_quantizers.40 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.99e-01 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.down_proj_weight_quantizers.41 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.down_proj_weight_quantizers.42 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.down_proj_weight_quantizers.43 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.down_proj_weight_quantizers.44 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.down_proj_weight_quantizers.45 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.down_proj_weight_quantizers.46 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.59e-02 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.down_proj_weight_quantizers.47 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.down_proj_weight_quantizers.48 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.down_proj_weight_quantizers.49 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.down_proj_weight_quantizers.50 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.down_proj_weight_quantizers.51 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.down_proj_weight_quantizers.52 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.17e-01 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.down_proj_weight_quantizers.53 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.down_proj_weight_quantizers.54 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.down_proj_weight_quantizers.55 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.down_proj_weight_quantizers.56 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.down_proj_weight_quantizers.57 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.down_proj_weight_quantizers.58 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.down_proj_weight_quantizers.59 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.down_proj_weight_quantizers.60 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.down_proj_weight_quantizers.61 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.down_proj_weight_quantizers.62 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.down_proj_weight_quantizers.63 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.down_proj_weight_quantizers.64 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.64e-02 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.down_proj_weight_quantizers.65 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.down_proj_weight_quantizers.66 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.down_proj_weight_quantizers.67 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.25e-01 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.down_proj_weight_quantizers.68 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.down_proj_weight_quantizers.69 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.64e-02 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.down_proj_weight_quantizers.70 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.50e-02 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.down_proj_weight_quantizers.71 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.down_proj_weight_quantizers.72 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.down_proj_weight_quantizers.73 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.down_proj_weight_quantizers.74 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.down_proj_weight_quantizers.75 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.down_proj_weight_quantizers.76 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.down_proj_weight_quantizers.77 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.54e-02 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.down_proj_weight_quantizers.78 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.27e-01 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.down_proj_weight_quantizers.79 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.down_proj_weight_quantizers.80 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.down_proj_weight_quantizers.81 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.25e-01 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.down_proj_weight_quantizers.82 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.down_proj_weight_quantizers.83 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.down_proj_weight_quantizers.84 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.down_proj_weight_quantizers.85 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.35e-01 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.down_proj_weight_quantizers.86 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.down_proj_weight_quantizers.87 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.17e-01 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.down_proj_weight_quantizers.88 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.64e-02 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.down_proj_weight_quantizers.89 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.59e-02 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.down_proj_weight_quantizers.90 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.17e-01 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.down_proj_weight_quantizers.91 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.down_proj_weight_quantizers.92 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.down_proj_weight_quantizers.93 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.46e-01 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.down_proj_weight_quantizers.94 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.60e-01 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.down_proj_weight_quantizers.95 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.64e-02 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.down_proj_weight_quantizers.96 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.down_proj_weight_quantizers.97 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.54e-02 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.down_proj_weight_quantizers.98 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.down_proj_weight_quantizers.99 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.down_proj_weight_quantizers.100 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.78e-01 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.down_proj_weight_quantizers.101 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.down_proj_weight_quantizers.102 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.down_proj_weight_quantizers.103 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.down_proj_weight_quantizers.104 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.down_proj_weight_quantizers.105 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.down_proj_weight_quantizers.106 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.50e-02 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.down_proj_weight_quantizers.107 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.45e-02 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.down_proj_weight_quantizers.108 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.down_proj_weight_quantizers.109 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.down_proj_weight_quantizers.110 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.down_proj_weight_quantizers.111 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.51e-01 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.down_proj_weight_quantizers.112 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.down_proj_weight_quantizers.113 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.down_proj_weight_quantizers.114 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.down_proj_weight_quantizers.115 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.down_proj_weight_quantizers.116 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.down_proj_weight_quantizers.117 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.down_proj_weight_quantizers.118 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.down_proj_weight_quantizers.119 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.down_proj_weight_quantizers.120 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.down_proj_weight_quantizers.121 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.down_proj_weight_quantizers.122 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.down_proj_weight_quantizers.123 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.down_proj_weight_quantizers.124 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.down_proj_weight_quantizers.125 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.41e-01 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.down_proj_weight_quantizers.126 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.31e-01 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.down_proj_weight_quantizers.127 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.54e-02 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.down_proj_weight_quantizers.128 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.down_proj_weight_quantizers.129 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.54e-02 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.down_proj_weight_quantizers.130 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.down_proj_weight_quantizers.131 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.down_proj_weight_quantizers.132 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.40e-02 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.down_proj_weight_quantizers.133 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.down_proj_weight_quantizers.134 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.down_proj_weight_quantizers.135 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.down_proj_weight_quantizers.136 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.down_proj_weight_quantizers.137 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.17e-01 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.down_proj_weight_quantizers.138 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.down_proj_weight_quantizers.139 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.down_proj_weight_quantizers.140 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.down_proj_weight_quantizers.141 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.down_proj_weight_quantizers.142 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.64e-02 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.down_proj_weight_quantizers.143 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.22e-01 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.down_proj_weight_quantizers.144 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.down_proj_weight_quantizers.145 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.down_proj_weight_quantizers.146 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.down_proj_weight_quantizers.147 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.down_proj_weight_quantizers.148 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.down_proj_weight_quantizers.149 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.down_proj_weight_quantizers.150 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.54e-02 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.down_proj_weight_quantizers.151 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.down_proj_weight_quantizers.152 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.down_proj_weight_quantizers.153 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.down_proj_weight_quantizers.154 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.down_proj_weight_quantizers.155 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.down_proj_weight_quantizers.156 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.64e-02 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.down_proj_weight_quantizers.157 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.down_proj_weight_quantizers.158 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.down_proj_weight_quantizers.159 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.down_proj_weight_quantizers.160 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.down_proj_weight_quantizers.161 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.down_proj_weight_quantizers.162 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.down_proj_weight_quantizers.163 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.down_proj_weight_quantizers.164 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.51e-01 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.down_proj_weight_quantizers.165 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.50e-01 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.down_proj_weight_quantizers.166 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.50e-02 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.down_proj_weight_quantizers.167 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.down_proj_weight_quantizers.168 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.down_proj_weight_quantizers.169 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.down_proj_weight_quantizers.170 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.down_proj_weight_quantizers.171 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.down_proj_weight_quantizers.172 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.46e-01 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.down_proj_weight_quantizers.173 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.22e-01 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.down_proj_weight_quantizers.174 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.down_proj_weight_quantizers.175 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.down_proj_weight_quantizers.176 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.down_proj_weight_quantizers.177 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.down_proj_weight_quantizers.178 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.down_proj_weight_quantizers.179 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.down_proj_weight_quantizers.180 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.down_proj_weight_quantizers.181 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.54e-02 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.down_proj_weight_quantizers.182 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.down_proj_weight_quantizers.183 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.down_proj_weight_quantizers.184 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.21e-01 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.down_proj_weight_quantizers.185 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.35e-02 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.down_proj_weight_quantizers.186 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.48e-01 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.down_proj_weight_quantizers.187 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.down_proj_weight_quantizers.188 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.down_proj_weight_quantizers.189 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.down_proj_weight_quantizers.190 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.down_proj_weight_quantizers.191 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.down_proj_weight_quantizers.192 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.down_proj_weight_quantizers.193 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.64e-02 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.down_proj_weight_quantizers.194 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.down_proj_weight_quantizers.195 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.down_proj_weight_quantizers.196 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.down_proj_weight_quantizers.197 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.down_proj_weight_quantizers.198 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.down_proj_weight_quantizers.199 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.down_proj_weight_quantizers.200 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.40e-02 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.down_proj_weight_quantizers.201 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.down_proj_weight_quantizers.202 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.15e-01 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.down_proj_weight_quantizers.203 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.down_proj_weight_quantizers.204 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.down_proj_weight_quantizers.205 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.62e-01 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.down_proj_weight_quantizers.206 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.down_proj_weight_quantizers.207 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.44e-01 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.down_proj_weight_quantizers.208 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.down_proj_weight_quantizers.209 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.04e-01 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.down_proj_weight_quantizers.210 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.down_proj_weight_quantizers.211 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.35e-01 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.down_proj_weight_quantizers.212 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.down_proj_weight_quantizers.213 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.26e-01 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.down_proj_weight_quantizers.214 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.down_proj_weight_quantizers.215 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.down_proj_weight_quantizers.216 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.down_proj_weight_quantizers.217 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.down_proj_weight_quantizers.218 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.45e-02 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.down_proj_weight_quantizers.219 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.down_proj_weight_quantizers.220 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.down_proj_weight_quantizers.221 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.down_proj_weight_quantizers.222 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.down_proj_weight_quantizers.223 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.29e-01 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.down_proj_weight_quantizers.224 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.down_proj_weight_quantizers.225 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.down_proj_weight_quantizers.226 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.down_proj_weight_quantizers.227 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.down_proj_weight_quantizers.228 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.down_proj_weight_quantizers.229 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.down_proj_weight_quantizers.230 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.down_proj_weight_quantizers.231 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.down_proj_weight_quantizers.232 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.down_proj_weight_quantizers.233 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.76e-01 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.down_proj_weight_quantizers.234 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.down_proj_weight_quantizers.235 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.64e-02 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.down_proj_weight_quantizers.236 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.down_proj_weight_quantizers.237 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.down_proj_weight_quantizers.238 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.down_proj_weight_quantizers.239 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.28e-01 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.down_proj_weight_quantizers.240 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.down_proj_weight_quantizers.241 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.down_proj_weight_quantizers.242 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.down_proj_weight_quantizers.243 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.down_proj_weight_quantizers.244 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.down_proj_weight_quantizers.245 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.down_proj_weight_quantizers.246 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.down_proj_weight_quantizers.247 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.down_proj_weight_quantizers.248 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.down_proj_weight_quantizers.249 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.down_proj_weight_quantizers.250 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.64e-02 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.down_proj_weight_quantizers.251 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.down_proj_weight_quantizers.252 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.down_proj_weight_quantizers.253 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.down_proj_weight_quantizers.254 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.down_proj_weight_quantizers.255 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.19.mlp.shared_experts.gate_proj.input_quantizer TensorQuantizer(disabled) +model.layers.19.mlp.shared_experts.gate_proj.output_quantizer TensorQuantizer(disabled) +model.layers.19.mlp.shared_experts.gate_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.19.mlp.shared_experts.up_proj.input_quantizer TensorQuantizer(disabled) +model.layers.19.mlp.shared_experts.up_proj.output_quantizer TensorQuantizer(disabled) +model.layers.19.mlp.shared_experts.up_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.19.mlp.shared_experts.down_proj.input_quantizer TensorQuantizer(disabled) +model.layers.19.mlp.shared_experts.down_proj.output_quantizer TensorQuantizer(disabled) +model.layers.19.mlp.shared_experts.down_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.20.self_attn.q_a_proj.input_quantizer TensorQuantizer(disabled) +model.layers.20.self_attn.q_a_proj.output_quantizer TensorQuantizer(disabled) +model.layers.20.self_attn.q_a_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.20.self_attn.q_b_proj.input_quantizer TensorQuantizer(disabled) +model.layers.20.self_attn.q_b_proj.output_quantizer TensorQuantizer(disabled) +model.layers.20.self_attn.q_b_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.20.self_attn.kv_a_proj_with_mqa.input_quantizer TensorQuantizer(disabled) +model.layers.20.self_attn.kv_a_proj_with_mqa.output_quantizer TensorQuantizer(disabled) +model.layers.20.self_attn.kv_a_proj_with_mqa.weight_quantizer TensorQuantizer(disabled) +model.layers.20.self_attn.kv_b_proj.input_quantizer TensorQuantizer(disabled) +model.layers.20.self_attn.kv_b_proj.output_quantizer TensorQuantizer(disabled) +model.layers.20.self_attn.kv_b_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.20.self_attn.o_proj.input_quantizer TensorQuantizer(disabled) +model.layers.20.self_attn.o_proj.output_quantizer TensorQuantizer(disabled) +model.layers.20.self_attn.o_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.20.self_attn.q_bmm_quantizer TensorQuantizer(disabled) +model.layers.20.self_attn.k_bmm_quantizer TensorQuantizer((4, 3) bit fake per-tensor amax=4.48e+02(const) calibrator=MaxCalibrator quant) +model.layers.20.self_attn.v_bmm_quantizer TensorQuantizer((4, 3) bit fake per-tensor amax=4.48e+02(const) calibrator=MaxCalibrator quant) +model.layers.20.self_attn.p_bmm_quantizer TensorQuantizer(disabled) +model.layers.20.mlp.experts.gate_up_proj_input_quantizer TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=7.77e-01 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.gate_up_proj_weight_quantizers.0 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.gate_up_proj_weight_quantizers.1 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.26e-01 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.gate_up_proj_weight_quantizers.2 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.gate_up_proj_weight_quantizers.3 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.33e-01 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.gate_up_proj_weight_quantizers.4 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.53e-01 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.gate_up_proj_weight_quantizers.5 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.gate_up_proj_weight_quantizers.6 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.gate_up_proj_weight_quantizers.7 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.gate_up_proj_weight_quantizers.8 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.gate_up_proj_weight_quantizers.9 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.gate_up_proj_weight_quantizers.10 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.15e-01 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.gate_up_proj_weight_quantizers.11 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.gate_up_proj_weight_quantizers.12 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.gate_up_proj_weight_quantizers.13 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.gate_up_proj_weight_quantizers.14 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.gate_up_proj_weight_quantizers.15 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.gate_up_proj_weight_quantizers.16 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.54e-01 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.gate_up_proj_weight_quantizers.17 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.gate_up_proj_weight_quantizers.18 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.gate_up_proj_weight_quantizers.19 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.gate_up_proj_weight_quantizers.20 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.20e-01 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.gate_up_proj_weight_quantizers.21 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.gate_up_proj_weight_quantizers.22 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.gate_up_proj_weight_quantizers.23 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.gate_up_proj_weight_quantizers.24 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=4.80e-01 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.gate_up_proj_weight_quantizers.25 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=4.18e-01 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.gate_up_proj_weight_quantizers.26 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.gate_up_proj_weight_quantizers.27 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.73e-01 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.gate_up_proj_weight_quantizers.28 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.gate_up_proj_weight_quantizers.29 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.89e-01 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.gate_up_proj_weight_quantizers.30 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.gate_up_proj_weight_quantizers.31 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.gate_up_proj_weight_quantizers.32 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.gate_up_proj_weight_quantizers.33 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.gate_up_proj_weight_quantizers.34 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.gate_up_proj_weight_quantizers.35 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.gate_up_proj_weight_quantizers.36 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.79e-01 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.gate_up_proj_weight_quantizers.37 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.gate_up_proj_weight_quantizers.38 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.gate_up_proj_weight_quantizers.39 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.gate_up_proj_weight_quantizers.40 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.gate_up_proj_weight_quantizers.41 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.gate_up_proj_weight_quantizers.42 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.gate_up_proj_weight_quantizers.43 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.gate_up_proj_weight_quantizers.44 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.37e-01 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.gate_up_proj_weight_quantizers.45 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.gate_up_proj_weight_quantizers.46 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.gate_up_proj_weight_quantizers.47 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.gate_up_proj_weight_quantizers.48 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.gate_up_proj_weight_quantizers.49 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.gate_up_proj_weight_quantizers.50 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.gate_up_proj_weight_quantizers.51 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.gate_up_proj_weight_quantizers.52 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.gate_up_proj_weight_quantizers.53 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.24e-01 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.gate_up_proj_weight_quantizers.54 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.gate_up_proj_weight_quantizers.55 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.gate_up_proj_weight_quantizers.56 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.gate_up_proj_weight_quantizers.57 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.gate_up_proj_weight_quantizers.58 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.gate_up_proj_weight_quantizers.59 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.gate_up_proj_weight_quantizers.60 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.gate_up_proj_weight_quantizers.61 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.gate_up_proj_weight_quantizers.62 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.gate_up_proj_weight_quantizers.63 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.24e-01 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.gate_up_proj_weight_quantizers.64 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.30e-01 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.gate_up_proj_weight_quantizers.65 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.gate_up_proj_weight_quantizers.66 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.gate_up_proj_weight_quantizers.67 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.gate_up_proj_weight_quantizers.68 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.gate_up_proj_weight_quantizers.69 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.32e-01 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.gate_up_proj_weight_quantizers.70 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.gate_up_proj_weight_quantizers.71 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.13e-01 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.gate_up_proj_weight_quantizers.72 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.gate_up_proj_weight_quantizers.73 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.gate_up_proj_weight_quantizers.74 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.gate_up_proj_weight_quantizers.75 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.gate_up_proj_weight_quantizers.76 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.gate_up_proj_weight_quantizers.77 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.gate_up_proj_weight_quantizers.78 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.gate_up_proj_weight_quantizers.79 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.gate_up_proj_weight_quantizers.80 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=5.78e-01 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.gate_up_proj_weight_quantizers.81 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.32e-01 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.gate_up_proj_weight_quantizers.82 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.22e-01 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.gate_up_proj_weight_quantizers.83 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.gate_up_proj_weight_quantizers.84 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.gate_up_proj_weight_quantizers.85 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.gate_up_proj_weight_quantizers.86 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.26e-01 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.gate_up_proj_weight_quantizers.87 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.gate_up_proj_weight_quantizers.88 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.gate_up_proj_weight_quantizers.89 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.18e-01 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.gate_up_proj_weight_quantizers.90 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.gate_up_proj_weight_quantizers.91 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.gate_up_proj_weight_quantizers.92 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.gate_up_proj_weight_quantizers.93 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.75e-01 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.gate_up_proj_weight_quantizers.94 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.gate_up_proj_weight_quantizers.95 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.gate_up_proj_weight_quantizers.96 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.gate_up_proj_weight_quantizers.97 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.gate_up_proj_weight_quantizers.98 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.gate_up_proj_weight_quantizers.99 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.gate_up_proj_weight_quantizers.100 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.gate_up_proj_weight_quantizers.101 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.45e-01 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.gate_up_proj_weight_quantizers.102 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.gate_up_proj_weight_quantizers.103 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.90e-01 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.gate_up_proj_weight_quantizers.104 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.gate_up_proj_weight_quantizers.105 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.gate_up_proj_weight_quantizers.106 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.gate_up_proj_weight_quantizers.107 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.gate_up_proj_weight_quantizers.108 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.gate_up_proj_weight_quantizers.109 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.77e-01 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.gate_up_proj_weight_quantizers.110 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.gate_up_proj_weight_quantizers.111 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.gate_up_proj_weight_quantizers.112 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.27e-01 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.gate_up_proj_weight_quantizers.113 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.gate_up_proj_weight_quantizers.114 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.gate_up_proj_weight_quantizers.115 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.14e-01 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.gate_up_proj_weight_quantizers.116 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.gate_up_proj_weight_quantizers.117 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.gate_up_proj_weight_quantizers.118 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.gate_up_proj_weight_quantizers.119 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.24e-01 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.gate_up_proj_weight_quantizers.120 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.08e-01 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.gate_up_proj_weight_quantizers.121 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.18e-01 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.gate_up_proj_weight_quantizers.122 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.gate_up_proj_weight_quantizers.123 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.gate_up_proj_weight_quantizers.124 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.gate_up_proj_weight_quantizers.125 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.gate_up_proj_weight_quantizers.126 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.gate_up_proj_weight_quantizers.127 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.gate_up_proj_weight_quantizers.128 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.41e-01 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.gate_up_proj_weight_quantizers.129 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.gate_up_proj_weight_quantizers.130 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.gate_up_proj_weight_quantizers.131 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.54e-01 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.gate_up_proj_weight_quantizers.132 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.gate_up_proj_weight_quantizers.133 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.68e-01 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.gate_up_proj_weight_quantizers.134 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.52e-01 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.gate_up_proj_weight_quantizers.135 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.gate_up_proj_weight_quantizers.136 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.gate_up_proj_weight_quantizers.137 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.gate_up_proj_weight_quantizers.138 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.gate_up_proj_weight_quantizers.139 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.gate_up_proj_weight_quantizers.140 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.gate_up_proj_weight_quantizers.141 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.gate_up_proj_weight_quantizers.142 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.gate_up_proj_weight_quantizers.143 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.gate_up_proj_weight_quantizers.144 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.gate_up_proj_weight_quantizers.145 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.gate_up_proj_weight_quantizers.146 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.gate_up_proj_weight_quantizers.147 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.gate_up_proj_weight_quantizers.148 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.gate_up_proj_weight_quantizers.149 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.gate_up_proj_weight_quantizers.150 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.gate_up_proj_weight_quantizers.151 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.gate_up_proj_weight_quantizers.152 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.gate_up_proj_weight_quantizers.153 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.15e-01 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.gate_up_proj_weight_quantizers.154 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.gate_up_proj_weight_quantizers.155 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.45e-01 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.gate_up_proj_weight_quantizers.156 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.gate_up_proj_weight_quantizers.157 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.33e-01 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.gate_up_proj_weight_quantizers.158 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.64e-02 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.gate_up_proj_weight_quantizers.159 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.57e-01 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.gate_up_proj_weight_quantizers.160 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.15e-01 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.gate_up_proj_weight_quantizers.161 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.gate_up_proj_weight_quantizers.162 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.gate_up_proj_weight_quantizers.163 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.gate_up_proj_weight_quantizers.164 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.gate_up_proj_weight_quantizers.165 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.gate_up_proj_weight_quantizers.166 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.gate_up_proj_weight_quantizers.167 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.gate_up_proj_weight_quantizers.168 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.gate_up_proj_weight_quantizers.169 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.gate_up_proj_weight_quantizers.170 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.gate_up_proj_weight_quantizers.171 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.gate_up_proj_weight_quantizers.172 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.53e-01 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.gate_up_proj_weight_quantizers.173 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.gate_up_proj_weight_quantizers.174 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.gate_up_proj_weight_quantizers.175 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.gate_up_proj_weight_quantizers.176 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.gate_up_proj_weight_quantizers.177 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.gate_up_proj_weight_quantizers.178 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.28e-01 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.gate_up_proj_weight_quantizers.179 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.gate_up_proj_weight_quantizers.180 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.gate_up_proj_weight_quantizers.181 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.gate_up_proj_weight_quantizers.182 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.gate_up_proj_weight_quantizers.183 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=7.03e-01 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.gate_up_proj_weight_quantizers.184 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.gate_up_proj_weight_quantizers.185 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.16e-01 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.gate_up_proj_weight_quantizers.186 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.gate_up_proj_weight_quantizers.187 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.gate_up_proj_weight_quantizers.188 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.gate_up_proj_weight_quantizers.189 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.gate_up_proj_weight_quantizers.190 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.gate_up_proj_weight_quantizers.191 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.gate_up_proj_weight_quantizers.192 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.37e-01 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.gate_up_proj_weight_quantizers.193 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.15e-01 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.gate_up_proj_weight_quantizers.194 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.66e-01 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.gate_up_proj_weight_quantizers.195 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.gate_up_proj_weight_quantizers.196 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.gate_up_proj_weight_quantizers.197 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.gate_up_proj_weight_quantizers.198 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.48e-01 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.gate_up_proj_weight_quantizers.199 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.gate_up_proj_weight_quantizers.200 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.gate_up_proj_weight_quantizers.201 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.gate_up_proj_weight_quantizers.202 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.gate_up_proj_weight_quantizers.203 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.gate_up_proj_weight_quantizers.204 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.54e-01 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.gate_up_proj_weight_quantizers.205 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.gate_up_proj_weight_quantizers.206 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.13e-01 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.gate_up_proj_weight_quantizers.207 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.gate_up_proj_weight_quantizers.208 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.gate_up_proj_weight_quantizers.209 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.gate_up_proj_weight_quantizers.210 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=4.08e-01 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.gate_up_proj_weight_quantizers.211 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.20e-01 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.gate_up_proj_weight_quantizers.212 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.gate_up_proj_weight_quantizers.213 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.gate_up_proj_weight_quantizers.214 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.gate_up_proj_weight_quantizers.215 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.41e-01 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.gate_up_proj_weight_quantizers.216 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.gate_up_proj_weight_quantizers.217 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.gate_up_proj_weight_quantizers.218 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.gate_up_proj_weight_quantizers.219 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.gate_up_proj_weight_quantizers.220 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.39e-01 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.gate_up_proj_weight_quantizers.221 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=4.61e-01 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.gate_up_proj_weight_quantizers.222 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.10e-01 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.gate_up_proj_weight_quantizers.223 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.gate_up_proj_weight_quantizers.224 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.gate_up_proj_weight_quantizers.225 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.25e-01 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.gate_up_proj_weight_quantizers.226 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.gate_up_proj_weight_quantizers.227 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.35e-01 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.gate_up_proj_weight_quantizers.228 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.gate_up_proj_weight_quantizers.229 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.gate_up_proj_weight_quantizers.230 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.gate_up_proj_weight_quantizers.231 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.63e-01 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.gate_up_proj_weight_quantizers.232 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.gate_up_proj_weight_quantizers.233 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.19e-01 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.gate_up_proj_weight_quantizers.234 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.16e-01 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.gate_up_proj_weight_quantizers.235 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.49e-01 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.gate_up_proj_weight_quantizers.236 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.gate_up_proj_weight_quantizers.237 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.gate_up_proj_weight_quantizers.238 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.gate_up_proj_weight_quantizers.239 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.gate_up_proj_weight_quantizers.240 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.gate_up_proj_weight_quantizers.241 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.gate_up_proj_weight_quantizers.242 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.gate_up_proj_weight_quantizers.243 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.41e-01 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.gate_up_proj_weight_quantizers.244 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.gate_up_proj_weight_quantizers.245 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.51e-01 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.gate_up_proj_weight_quantizers.246 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.gate_up_proj_weight_quantizers.247 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.14e-01 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.gate_up_proj_weight_quantizers.248 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.gate_up_proj_weight_quantizers.249 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.gate_up_proj_weight_quantizers.250 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.gate_up_proj_weight_quantizers.251 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.gate_up_proj_weight_quantizers.252 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.gate_up_proj_weight_quantizers.253 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.gate_up_proj_weight_quantizers.254 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.gate_up_proj_weight_quantizers.255 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.down_proj_input_quantizer TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e+01 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.down_proj_weight_quantizers.0 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.down_proj_weight_quantizers.1 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.down_proj_weight_quantizers.2 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.down_proj_weight_quantizers.3 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.14e-01 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.down_proj_weight_quantizers.4 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.65e-01 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.down_proj_weight_quantizers.5 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.down_proj_weight_quantizers.6 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.down_proj_weight_quantizers.7 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.down_proj_weight_quantizers.8 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.down_proj_weight_quantizers.9 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.down_proj_weight_quantizers.10 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.down_proj_weight_quantizers.11 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.down_proj_weight_quantizers.12 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.down_proj_weight_quantizers.13 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.down_proj_weight_quantizers.14 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.down_proj_weight_quantizers.15 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.down_proj_weight_quantizers.16 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.down_proj_weight_quantizers.17 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.down_proj_weight_quantizers.18 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.down_proj_weight_quantizers.19 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.down_proj_weight_quantizers.20 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.29e-01 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.down_proj_weight_quantizers.21 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.down_proj_weight_quantizers.22 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.50e-02 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.down_proj_weight_quantizers.23 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.down_proj_weight_quantizers.24 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.down_proj_weight_quantizers.25 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.down_proj_weight_quantizers.26 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.down_proj_weight_quantizers.27 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.down_proj_weight_quantizers.28 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.down_proj_weight_quantizers.29 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.57e-01 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.down_proj_weight_quantizers.30 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.down_proj_weight_quantizers.31 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.down_proj_weight_quantizers.32 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.down_proj_weight_quantizers.33 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.down_proj_weight_quantizers.34 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.44e-01 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.down_proj_weight_quantizers.35 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.54e-02 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.down_proj_weight_quantizers.36 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.down_proj_weight_quantizers.37 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.down_proj_weight_quantizers.38 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.59e-02 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.down_proj_weight_quantizers.39 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.down_proj_weight_quantizers.40 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.64e-02 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.down_proj_weight_quantizers.41 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.21e-01 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.down_proj_weight_quantizers.42 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.64e-02 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.down_proj_weight_quantizers.43 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.down_proj_weight_quantizers.44 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.down_proj_weight_quantizers.45 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.35e-02 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.down_proj_weight_quantizers.46 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.down_proj_weight_quantizers.47 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.down_proj_weight_quantizers.48 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.50e-02 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.down_proj_weight_quantizers.49 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.down_proj_weight_quantizers.50 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.down_proj_weight_quantizers.51 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.down_proj_weight_quantizers.52 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.79e-01 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.down_proj_weight_quantizers.53 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.38e-01 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.down_proj_weight_quantizers.54 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.54e-02 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.down_proj_weight_quantizers.55 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.54e-02 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.down_proj_weight_quantizers.56 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.down_proj_weight_quantizers.57 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.down_proj_weight_quantizers.58 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.down_proj_weight_quantizers.59 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.25e-01 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.down_proj_weight_quantizers.60 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.down_proj_weight_quantizers.61 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.down_proj_weight_quantizers.62 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.down_proj_weight_quantizers.63 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.down_proj_weight_quantizers.64 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.67e-01 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.down_proj_weight_quantizers.65 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.down_proj_weight_quantizers.66 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.down_proj_weight_quantizers.67 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.down_proj_weight_quantizers.68 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.down_proj_weight_quantizers.69 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.33e-01 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.down_proj_weight_quantizers.70 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.down_proj_weight_quantizers.71 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.down_proj_weight_quantizers.72 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.down_proj_weight_quantizers.73 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.64e-02 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.down_proj_weight_quantizers.74 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.64e-02 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.down_proj_weight_quantizers.75 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.down_proj_weight_quantizers.76 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.down_proj_weight_quantizers.77 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.down_proj_weight_quantizers.78 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.down_proj_weight_quantizers.79 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.down_proj_weight_quantizers.80 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.16e-01 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.down_proj_weight_quantizers.81 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.down_proj_weight_quantizers.82 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.64e-01 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.down_proj_weight_quantizers.83 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.down_proj_weight_quantizers.84 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.down_proj_weight_quantizers.85 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.down_proj_weight_quantizers.86 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.down_proj_weight_quantizers.87 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.25e-01 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.down_proj_weight_quantizers.88 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.down_proj_weight_quantizers.89 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.down_proj_weight_quantizers.90 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.down_proj_weight_quantizers.91 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.down_proj_weight_quantizers.92 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.down_proj_weight_quantizers.93 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.28e-01 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.down_proj_weight_quantizers.94 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.down_proj_weight_quantizers.95 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.35e-02 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.down_proj_weight_quantizers.96 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.down_proj_weight_quantizers.97 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.down_proj_weight_quantizers.98 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.down_proj_weight_quantizers.99 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.45e-02 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.down_proj_weight_quantizers.100 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.73e-01 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.down_proj_weight_quantizers.101 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.down_proj_weight_quantizers.102 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.down_proj_weight_quantizers.103 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.down_proj_weight_quantizers.104 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.down_proj_weight_quantizers.105 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.down_proj_weight_quantizers.106 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.down_proj_weight_quantizers.107 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.down_proj_weight_quantizers.108 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.down_proj_weight_quantizers.109 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.37e-01 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.down_proj_weight_quantizers.110 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.down_proj_weight_quantizers.111 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.down_proj_weight_quantizers.112 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.41e-01 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.down_proj_weight_quantizers.113 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.down_proj_weight_quantizers.114 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.down_proj_weight_quantizers.115 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.down_proj_weight_quantizers.116 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.down_proj_weight_quantizers.117 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.40e-02 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.down_proj_weight_quantizers.118 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.down_proj_weight_quantizers.119 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.down_proj_weight_quantizers.120 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.46e-01 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.down_proj_weight_quantizers.121 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.down_proj_weight_quantizers.122 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.31e-01 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.down_proj_weight_quantizers.123 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.down_proj_weight_quantizers.124 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.down_proj_weight_quantizers.125 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.50e-02 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.down_proj_weight_quantizers.126 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.down_proj_weight_quantizers.127 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.down_proj_weight_quantizers.128 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.67e-01 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.down_proj_weight_quantizers.129 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.down_proj_weight_quantizers.130 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.down_proj_weight_quantizers.131 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.47e-01 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.down_proj_weight_quantizers.132 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.64e-02 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.down_proj_weight_quantizers.133 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.15e-01 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.down_proj_weight_quantizers.134 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.15e-01 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.down_proj_weight_quantizers.135 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.down_proj_weight_quantizers.136 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.down_proj_weight_quantizers.137 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.down_proj_weight_quantizers.138 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.54e-02 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.down_proj_weight_quantizers.139 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.down_proj_weight_quantizers.140 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.down_proj_weight_quantizers.141 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.down_proj_weight_quantizers.142 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.down_proj_weight_quantizers.143 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.down_proj_weight_quantizers.144 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.down_proj_weight_quantizers.145 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.down_proj_weight_quantizers.146 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.64e-02 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.down_proj_weight_quantizers.147 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.40e-02 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.down_proj_weight_quantizers.148 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.15e-01 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.down_proj_weight_quantizers.149 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.down_proj_weight_quantizers.150 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.down_proj_weight_quantizers.151 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.down_proj_weight_quantizers.152 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.64e-02 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.down_proj_weight_quantizers.153 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.down_proj_weight_quantizers.154 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.down_proj_weight_quantizers.155 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.24e-01 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.down_proj_weight_quantizers.156 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.down_proj_weight_quantizers.157 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.25e-01 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.down_proj_weight_quantizers.158 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.64e-02 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.down_proj_weight_quantizers.159 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.down_proj_weight_quantizers.160 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.50e-02 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.down_proj_weight_quantizers.161 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.down_proj_weight_quantizers.162 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.down_proj_weight_quantizers.163 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.down_proj_weight_quantizers.164 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.down_proj_weight_quantizers.165 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.40e-02 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.down_proj_weight_quantizers.166 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.down_proj_weight_quantizers.167 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.down_proj_weight_quantizers.168 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.down_proj_weight_quantizers.169 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.down_proj_weight_quantizers.170 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.54e-02 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.down_proj_weight_quantizers.171 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.28e-01 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.down_proj_weight_quantizers.172 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.64e-02 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.down_proj_weight_quantizers.173 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.down_proj_weight_quantizers.174 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.down_proj_weight_quantizers.175 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.down_proj_weight_quantizers.176 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.down_proj_weight_quantizers.177 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.down_proj_weight_quantizers.178 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.down_proj_weight_quantizers.179 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.down_proj_weight_quantizers.180 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.45e-02 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.down_proj_weight_quantizers.181 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.down_proj_weight_quantizers.182 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.down_proj_weight_quantizers.183 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.24e-01 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.down_proj_weight_quantizers.184 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.down_proj_weight_quantizers.185 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.81e-01 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.down_proj_weight_quantizers.186 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.down_proj_weight_quantizers.187 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.down_proj_weight_quantizers.188 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.down_proj_weight_quantizers.189 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.down_proj_weight_quantizers.190 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.down_proj_weight_quantizers.191 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.54e-02 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.down_proj_weight_quantizers.192 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.down_proj_weight_quantizers.193 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.14e-01 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.down_proj_weight_quantizers.194 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.59e-01 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.down_proj_weight_quantizers.195 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.down_proj_weight_quantizers.196 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.54e-02 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.down_proj_weight_quantizers.197 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.59e-02 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.down_proj_weight_quantizers.198 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.70e-01 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.down_proj_weight_quantizers.199 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.down_proj_weight_quantizers.200 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.64e-02 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.down_proj_weight_quantizers.201 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.down_proj_weight_quantizers.202 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.down_proj_weight_quantizers.203 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.down_proj_weight_quantizers.204 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.down_proj_weight_quantizers.205 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.down_proj_weight_quantizers.206 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.down_proj_weight_quantizers.207 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.down_proj_weight_quantizers.208 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.down_proj_weight_quantizers.209 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.down_proj_weight_quantizers.210 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.64e-01 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.down_proj_weight_quantizers.211 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.down_proj_weight_quantizers.212 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.down_proj_weight_quantizers.213 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.down_proj_weight_quantizers.214 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.down_proj_weight_quantizers.215 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.40e-01 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.down_proj_weight_quantizers.216 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.down_proj_weight_quantizers.217 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.down_proj_weight_quantizers.218 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.down_proj_weight_quantizers.219 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.down_proj_weight_quantizers.220 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.down_proj_weight_quantizers.221 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.20e-01 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.down_proj_weight_quantizers.222 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.48e-01 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.down_proj_weight_quantizers.223 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.54e-02 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.down_proj_weight_quantizers.224 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.64e-02 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.down_proj_weight_quantizers.225 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.down_proj_weight_quantizers.226 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.64e-02 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.down_proj_weight_quantizers.227 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.15e-01 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.down_proj_weight_quantizers.228 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.down_proj_weight_quantizers.229 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.down_proj_weight_quantizers.230 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.24e-01 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.down_proj_weight_quantizers.231 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.down_proj_weight_quantizers.232 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.down_proj_weight_quantizers.233 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.down_proj_weight_quantizers.234 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.24e-01 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.down_proj_weight_quantizers.235 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.27e-01 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.down_proj_weight_quantizers.236 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.down_proj_weight_quantizers.237 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.down_proj_weight_quantizers.238 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.down_proj_weight_quantizers.239 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.down_proj_weight_quantizers.240 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.45e-02 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.down_proj_weight_quantizers.241 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.down_proj_weight_quantizers.242 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.down_proj_weight_quantizers.243 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.35e-01 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.down_proj_weight_quantizers.244 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.down_proj_weight_quantizers.245 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.down_proj_weight_quantizers.246 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.down_proj_weight_quantizers.247 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.23e-01 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.down_proj_weight_quantizers.248 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.down_proj_weight_quantizers.249 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.down_proj_weight_quantizers.250 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.down_proj_weight_quantizers.251 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.down_proj_weight_quantizers.252 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.down_proj_weight_quantizers.253 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.down_proj_weight_quantizers.254 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.down_proj_weight_quantizers.255 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.20.mlp.shared_experts.gate_proj.input_quantizer TensorQuantizer(disabled) +model.layers.20.mlp.shared_experts.gate_proj.output_quantizer TensorQuantizer(disabled) +model.layers.20.mlp.shared_experts.gate_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.20.mlp.shared_experts.up_proj.input_quantizer TensorQuantizer(disabled) +model.layers.20.mlp.shared_experts.up_proj.output_quantizer TensorQuantizer(disabled) +model.layers.20.mlp.shared_experts.up_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.20.mlp.shared_experts.down_proj.input_quantizer TensorQuantizer(disabled) +model.layers.20.mlp.shared_experts.down_proj.output_quantizer TensorQuantizer(disabled) +model.layers.20.mlp.shared_experts.down_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.21.self_attn.q_a_proj.input_quantizer TensorQuantizer(disabled) +model.layers.21.self_attn.q_a_proj.output_quantizer TensorQuantizer(disabled) +model.layers.21.self_attn.q_a_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.21.self_attn.q_b_proj.input_quantizer TensorQuantizer(disabled) +model.layers.21.self_attn.q_b_proj.output_quantizer TensorQuantizer(disabled) +model.layers.21.self_attn.q_b_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.21.self_attn.kv_a_proj_with_mqa.input_quantizer TensorQuantizer(disabled) +model.layers.21.self_attn.kv_a_proj_with_mqa.output_quantizer TensorQuantizer(disabled) +model.layers.21.self_attn.kv_a_proj_with_mqa.weight_quantizer TensorQuantizer(disabled) +model.layers.21.self_attn.kv_b_proj.input_quantizer TensorQuantizer(disabled) +model.layers.21.self_attn.kv_b_proj.output_quantizer TensorQuantizer(disabled) +model.layers.21.self_attn.kv_b_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.21.self_attn.o_proj.input_quantizer TensorQuantizer(disabled) +model.layers.21.self_attn.o_proj.output_quantizer TensorQuantizer(disabled) +model.layers.21.self_attn.o_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.21.self_attn.q_bmm_quantizer TensorQuantizer(disabled) +model.layers.21.self_attn.k_bmm_quantizer TensorQuantizer((4, 3) bit fake per-tensor amax=4.48e+02(const) calibrator=MaxCalibrator quant) +model.layers.21.self_attn.v_bmm_quantizer TensorQuantizer((4, 3) bit fake per-tensor amax=4.48e+02(const) calibrator=MaxCalibrator quant) +model.layers.21.self_attn.p_bmm_quantizer TensorQuantizer(disabled) +model.layers.21.mlp.experts.gate_up_proj_input_quantizer TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=7.89e-01 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.gate_up_proj_weight_quantizers.0 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.gate_up_proj_weight_quantizers.1 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.gate_up_proj_weight_quantizers.2 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.26e-01 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.gate_up_proj_weight_quantizers.3 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.gate_up_proj_weight_quantizers.4 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.gate_up_proj_weight_quantizers.5 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.25e-01 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.gate_up_proj_weight_quantizers.6 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.25e-01 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.gate_up_proj_weight_quantizers.7 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.gate_up_proj_weight_quantizers.8 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.gate_up_proj_weight_quantizers.9 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.gate_up_proj_weight_quantizers.10 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.gate_up_proj_weight_quantizers.11 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.23e-01 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.gate_up_proj_weight_quantizers.12 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=4.00e-01 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.gate_up_proj_weight_quantizers.13 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.gate_up_proj_weight_quantizers.14 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.37e-01 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.gate_up_proj_weight_quantizers.15 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.48e-01 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.gate_up_proj_weight_quantizers.16 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.gate_up_proj_weight_quantizers.17 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.gate_up_proj_weight_quantizers.18 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.gate_up_proj_weight_quantizers.19 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.91e-01 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.gate_up_proj_weight_quantizers.20 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.gate_up_proj_weight_quantizers.21 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.gate_up_proj_weight_quantizers.22 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.gate_up_proj_weight_quantizers.23 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.gate_up_proj_weight_quantizers.24 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.gate_up_proj_weight_quantizers.25 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.58e-01 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.gate_up_proj_weight_quantizers.26 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.gate_up_proj_weight_quantizers.27 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.gate_up_proj_weight_quantizers.28 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.gate_up_proj_weight_quantizers.29 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.gate_up_proj_weight_quantizers.30 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.88e-01 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.gate_up_proj_weight_quantizers.31 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.45e-02 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.gate_up_proj_weight_quantizers.32 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.gate_up_proj_weight_quantizers.33 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.gate_up_proj_weight_quantizers.34 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.gate_up_proj_weight_quantizers.35 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.gate_up_proj_weight_quantizers.36 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.gate_up_proj_weight_quantizers.37 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.gate_up_proj_weight_quantizers.38 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.gate_up_proj_weight_quantizers.39 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.gate_up_proj_weight_quantizers.40 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.gate_up_proj_weight_quantizers.41 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.gate_up_proj_weight_quantizers.42 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.37e-01 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.gate_up_proj_weight_quantizers.43 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.gate_up_proj_weight_quantizers.44 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.gate_up_proj_weight_quantizers.45 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.gate_up_proj_weight_quantizers.46 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.gate_up_proj_weight_quantizers.47 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.gate_up_proj_weight_quantizers.48 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.gate_up_proj_weight_quantizers.49 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.gate_up_proj_weight_quantizers.50 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.19e-01 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.gate_up_proj_weight_quantizers.51 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.46e-01 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.gate_up_proj_weight_quantizers.52 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.gate_up_proj_weight_quantizers.53 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.gate_up_proj_weight_quantizers.54 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.gate_up_proj_weight_quantizers.55 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.gate_up_proj_weight_quantizers.56 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.gate_up_proj_weight_quantizers.57 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.gate_up_proj_weight_quantizers.58 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.gate_up_proj_weight_quantizers.59 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.gate_up_proj_weight_quantizers.60 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.gate_up_proj_weight_quantizers.61 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.gate_up_proj_weight_quantizers.62 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.gate_up_proj_weight_quantizers.63 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.gate_up_proj_weight_quantizers.64 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.gate_up_proj_weight_quantizers.65 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.gate_up_proj_weight_quantizers.66 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.gate_up_proj_weight_quantizers.67 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.gate_up_proj_weight_quantizers.68 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.gate_up_proj_weight_quantizers.69 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.gate_up_proj_weight_quantizers.70 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.gate_up_proj_weight_quantizers.71 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.63e-01 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.gate_up_proj_weight_quantizers.72 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.91e-01 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.gate_up_proj_weight_quantizers.73 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.88e-01 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.gate_up_proj_weight_quantizers.74 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.gate_up_proj_weight_quantizers.75 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.gate_up_proj_weight_quantizers.76 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.gate_up_proj_weight_quantizers.77 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.gate_up_proj_weight_quantizers.78 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.gate_up_proj_weight_quantizers.79 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.gate_up_proj_weight_quantizers.80 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.gate_up_proj_weight_quantizers.81 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.gate_up_proj_weight_quantizers.82 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.44e-01 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.gate_up_proj_weight_quantizers.83 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.gate_up_proj_weight_quantizers.84 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.57e-01 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.gate_up_proj_weight_quantizers.85 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.gate_up_proj_weight_quantizers.86 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.gate_up_proj_weight_quantizers.87 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.gate_up_proj_weight_quantizers.88 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.gate_up_proj_weight_quantizers.89 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.76e-01 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.gate_up_proj_weight_quantizers.90 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.gate_up_proj_weight_quantizers.91 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.gate_up_proj_weight_quantizers.92 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.30e-01 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.gate_up_proj_weight_quantizers.93 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.34e-01 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.gate_up_proj_weight_quantizers.94 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.gate_up_proj_weight_quantizers.95 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.gate_up_proj_weight_quantizers.96 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.gate_up_proj_weight_quantizers.97 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.gate_up_proj_weight_quantizers.98 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.gate_up_proj_weight_quantizers.99 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.15e-01 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.gate_up_proj_weight_quantizers.100 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.01e-01 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.gate_up_proj_weight_quantizers.101 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.gate_up_proj_weight_quantizers.102 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.13e-01 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.gate_up_proj_weight_quantizers.103 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.gate_up_proj_weight_quantizers.104 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.57e-01 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.gate_up_proj_weight_quantizers.105 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.gate_up_proj_weight_quantizers.106 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.gate_up_proj_weight_quantizers.107 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.11e-01 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.gate_up_proj_weight_quantizers.108 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.gate_up_proj_weight_quantizers.109 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.gate_up_proj_weight_quantizers.110 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.gate_up_proj_weight_quantizers.111 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.gate_up_proj_weight_quantizers.112 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.gate_up_proj_weight_quantizers.113 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.23e-01 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.gate_up_proj_weight_quantizers.114 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.gate_up_proj_weight_quantizers.115 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.32e-01 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.gate_up_proj_weight_quantizers.116 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.gate_up_proj_weight_quantizers.117 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.gate_up_proj_weight_quantizers.118 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.gate_up_proj_weight_quantizers.119 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.gate_up_proj_weight_quantizers.120 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.gate_up_proj_weight_quantizers.121 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.54e-01 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.gate_up_proj_weight_quantizers.122 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.gate_up_proj_weight_quantizers.123 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.gate_up_proj_weight_quantizers.124 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.gate_up_proj_weight_quantizers.125 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.gate_up_proj_weight_quantizers.126 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.14e-01 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.gate_up_proj_weight_quantizers.127 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.36e-01 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.gate_up_proj_weight_quantizers.128 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.gate_up_proj_weight_quantizers.129 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.gate_up_proj_weight_quantizers.130 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.73e-01 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.gate_up_proj_weight_quantizers.131 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.gate_up_proj_weight_quantizers.132 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.gate_up_proj_weight_quantizers.133 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.31e-01 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.gate_up_proj_weight_quantizers.134 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.gate_up_proj_weight_quantizers.135 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.gate_up_proj_weight_quantizers.136 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.gate_up_proj_weight_quantizers.137 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.01e-01 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.gate_up_proj_weight_quantizers.138 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.gate_up_proj_weight_quantizers.139 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.gate_up_proj_weight_quantizers.140 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.20e-01 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.gate_up_proj_weight_quantizers.141 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.64e-02 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.gate_up_proj_weight_quantizers.142 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.gate_up_proj_weight_quantizers.143 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.gate_up_proj_weight_quantizers.144 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.gate_up_proj_weight_quantizers.145 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.gate_up_proj_weight_quantizers.146 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.gate_up_proj_weight_quantizers.147 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.gate_up_proj_weight_quantizers.148 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.65e-01 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.gate_up_proj_weight_quantizers.149 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.gate_up_proj_weight_quantizers.150 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.gate_up_proj_weight_quantizers.151 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.gate_up_proj_weight_quantizers.152 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.gate_up_proj_weight_quantizers.153 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.gate_up_proj_weight_quantizers.154 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.gate_up_proj_weight_quantizers.155 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.gate_up_proj_weight_quantizers.156 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.81e-01 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.gate_up_proj_weight_quantizers.157 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.gate_up_proj_weight_quantizers.158 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.gate_up_proj_weight_quantizers.159 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.gate_up_proj_weight_quantizers.160 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.gate_up_proj_weight_quantizers.161 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.gate_up_proj_weight_quantizers.162 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.gate_up_proj_weight_quantizers.163 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.gate_up_proj_weight_quantizers.164 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.gate_up_proj_weight_quantizers.165 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.gate_up_proj_weight_quantizers.166 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.gate_up_proj_weight_quantizers.167 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.gate_up_proj_weight_quantizers.168 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.gate_up_proj_weight_quantizers.169 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.gate_up_proj_weight_quantizers.170 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.gate_up_proj_weight_quantizers.171 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.gate_up_proj_weight_quantizers.172 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.gate_up_proj_weight_quantizers.173 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.gate_up_proj_weight_quantizers.174 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.gate_up_proj_weight_quantizers.175 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.gate_up_proj_weight_quantizers.176 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.gate_up_proj_weight_quantizers.177 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.gate_up_proj_weight_quantizers.178 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.gate_up_proj_weight_quantizers.179 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.gate_up_proj_weight_quantizers.180 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.gate_up_proj_weight_quantizers.181 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.gate_up_proj_weight_quantizers.182 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.gate_up_proj_weight_quantizers.183 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.gate_up_proj_weight_quantizers.184 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.64e-02 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.gate_up_proj_weight_quantizers.185 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.gate_up_proj_weight_quantizers.186 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.34e-01 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.gate_up_proj_weight_quantizers.187 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.gate_up_proj_weight_quantizers.188 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.gate_up_proj_weight_quantizers.189 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.15e-01 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.gate_up_proj_weight_quantizers.190 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.gate_up_proj_weight_quantizers.191 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.gate_up_proj_weight_quantizers.192 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.gate_up_proj_weight_quantizers.193 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.gate_up_proj_weight_quantizers.194 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.gate_up_proj_weight_quantizers.195 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.gate_up_proj_weight_quantizers.196 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.gate_up_proj_weight_quantizers.197 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.gate_up_proj_weight_quantizers.198 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.gate_up_proj_weight_quantizers.199 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.gate_up_proj_weight_quantizers.200 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.gate_up_proj_weight_quantizers.201 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.gate_up_proj_weight_quantizers.202 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.25e-01 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.gate_up_proj_weight_quantizers.203 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.gate_up_proj_weight_quantizers.204 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.gate_up_proj_weight_quantizers.205 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.36e-01 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.gate_up_proj_weight_quantizers.206 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.gate_up_proj_weight_quantizers.207 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.gate_up_proj_weight_quantizers.208 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.gate_up_proj_weight_quantizers.209 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.gate_up_proj_weight_quantizers.210 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.45e-02 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.gate_up_proj_weight_quantizers.211 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.gate_up_proj_weight_quantizers.212 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.gate_up_proj_weight_quantizers.213 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.gate_up_proj_weight_quantizers.214 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.88e-01 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.gate_up_proj_weight_quantizers.215 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.gate_up_proj_weight_quantizers.216 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.33e-01 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.gate_up_proj_weight_quantizers.217 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.gate_up_proj_weight_quantizers.218 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.gate_up_proj_weight_quantizers.219 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.gate_up_proj_weight_quantizers.220 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.gate_up_proj_weight_quantizers.221 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.gate_up_proj_weight_quantizers.222 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.gate_up_proj_weight_quantizers.223 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.gate_up_proj_weight_quantizers.224 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.gate_up_proj_weight_quantizers.225 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.gate_up_proj_weight_quantizers.226 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.gate_up_proj_weight_quantizers.227 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.gate_up_proj_weight_quantizers.228 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.gate_up_proj_weight_quantizers.229 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.gate_up_proj_weight_quantizers.230 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.gate_up_proj_weight_quantizers.231 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.14e-01 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.gate_up_proj_weight_quantizers.232 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.gate_up_proj_weight_quantizers.233 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.gate_up_proj_weight_quantizers.234 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.gate_up_proj_weight_quantizers.235 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.gate_up_proj_weight_quantizers.236 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.gate_up_proj_weight_quantizers.237 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.gate_up_proj_weight_quantizers.238 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.gate_up_proj_weight_quantizers.239 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.gate_up_proj_weight_quantizers.240 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.gate_up_proj_weight_quantizers.241 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.gate_up_proj_weight_quantizers.242 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.gate_up_proj_weight_quantizers.243 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.gate_up_proj_weight_quantizers.244 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.31e-01 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.gate_up_proj_weight_quantizers.245 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.gate_up_proj_weight_quantizers.246 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.gate_up_proj_weight_quantizers.247 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.gate_up_proj_weight_quantizers.248 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.gate_up_proj_weight_quantizers.249 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.gate_up_proj_weight_quantizers.250 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.gate_up_proj_weight_quantizers.251 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.gate_up_proj_weight_quantizers.252 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.gate_up_proj_weight_quantizers.253 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.82e-01 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.gate_up_proj_weight_quantizers.254 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.gate_up_proj_weight_quantizers.255 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.down_proj_input_quantizer TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=6.06e+00 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.down_proj_weight_quantizers.0 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.41e-01 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.down_proj_weight_quantizers.1 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.down_proj_weight_quantizers.2 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.down_proj_weight_quantizers.3 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.down_proj_weight_quantizers.4 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.down_proj_weight_quantizers.5 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.down_proj_weight_quantizers.6 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.down_proj_weight_quantizers.7 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.down_proj_weight_quantizers.8 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.down_proj_weight_quantizers.9 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.down_proj_weight_quantizers.10 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.25e-02 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.down_proj_weight_quantizers.11 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.19e-01 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.down_proj_weight_quantizers.12 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.35e-01 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.down_proj_weight_quantizers.13 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.down_proj_weight_quantizers.14 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.87e-01 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.down_proj_weight_quantizers.15 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.down_proj_weight_quantizers.16 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.down_proj_weight_quantizers.17 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.down_proj_weight_quantizers.18 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.down_proj_weight_quantizers.19 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.62e-01 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.down_proj_weight_quantizers.20 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.down_proj_weight_quantizers.21 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.down_proj_weight_quantizers.22 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.down_proj_weight_quantizers.23 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.16e-01 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.down_proj_weight_quantizers.24 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.59e-02 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.down_proj_weight_quantizers.25 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.37e-01 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.down_proj_weight_quantizers.26 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.down_proj_weight_quantizers.27 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.down_proj_weight_quantizers.28 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.down_proj_weight_quantizers.29 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.down_proj_weight_quantizers.30 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.28e-01 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.down_proj_weight_quantizers.31 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.down_proj_weight_quantizers.32 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.down_proj_weight_quantizers.33 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.down_proj_weight_quantizers.34 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.down_proj_weight_quantizers.35 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.down_proj_weight_quantizers.36 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.down_proj_weight_quantizers.37 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.down_proj_weight_quantizers.38 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.59e-02 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.down_proj_weight_quantizers.39 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.down_proj_weight_quantizers.40 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.down_proj_weight_quantizers.41 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.down_proj_weight_quantizers.42 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.down_proj_weight_quantizers.43 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.54e-02 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.down_proj_weight_quantizers.44 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.54e-02 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.down_proj_weight_quantizers.45 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.down_proj_weight_quantizers.46 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.down_proj_weight_quantizers.47 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.down_proj_weight_quantizers.48 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.down_proj_weight_quantizers.49 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.64e-02 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.down_proj_weight_quantizers.50 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.down_proj_weight_quantizers.51 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.16e-01 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.down_proj_weight_quantizers.52 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.54e-02 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.down_proj_weight_quantizers.53 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.down_proj_weight_quantizers.54 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.64e-02 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.down_proj_weight_quantizers.55 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.50e-02 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.down_proj_weight_quantizers.56 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.down_proj_weight_quantizers.57 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.46e-01 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.down_proj_weight_quantizers.58 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.down_proj_weight_quantizers.59 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.down_proj_weight_quantizers.60 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.down_proj_weight_quantizers.61 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.down_proj_weight_quantizers.62 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.64e-02 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.down_proj_weight_quantizers.63 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.50e-02 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.down_proj_weight_quantizers.64 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.down_proj_weight_quantizers.65 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.down_proj_weight_quantizers.66 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.50e-02 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.down_proj_weight_quantizers.67 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.25e-01 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.down_proj_weight_quantizers.68 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.down_proj_weight_quantizers.69 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.59e-02 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.down_proj_weight_quantizers.70 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.40e-02 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.down_proj_weight_quantizers.71 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.down_proj_weight_quantizers.72 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.44e-01 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.down_proj_weight_quantizers.73 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.down_proj_weight_quantizers.74 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.down_proj_weight_quantizers.75 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.29e-01 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.down_proj_weight_quantizers.76 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.64e-02 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.down_proj_weight_quantizers.77 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.down_proj_weight_quantizers.78 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.down_proj_weight_quantizers.79 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.down_proj_weight_quantizers.80 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.down_proj_weight_quantizers.81 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.down_proj_weight_quantizers.82 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.95e-01 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.down_proj_weight_quantizers.83 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.down_proj_weight_quantizers.84 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.45e-01 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.down_proj_weight_quantizers.85 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.down_proj_weight_quantizers.86 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.down_proj_weight_quantizers.87 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.down_proj_weight_quantizers.88 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.down_proj_weight_quantizers.89 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.20e-01 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.down_proj_weight_quantizers.90 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.down_proj_weight_quantizers.91 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.down_proj_weight_quantizers.92 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.down_proj_weight_quantizers.93 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.down_proj_weight_quantizers.94 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.64e-02 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.down_proj_weight_quantizers.95 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.27e-01 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.down_proj_weight_quantizers.96 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.down_proj_weight_quantizers.97 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.down_proj_weight_quantizers.98 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.35e-02 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.down_proj_weight_quantizers.99 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.down_proj_weight_quantizers.100 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.19e-01 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.down_proj_weight_quantizers.101 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.down_proj_weight_quantizers.102 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.down_proj_weight_quantizers.103 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.down_proj_weight_quantizers.104 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.27e-01 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.down_proj_weight_quantizers.105 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.45e-02 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.down_proj_weight_quantizers.106 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.down_proj_weight_quantizers.107 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.down_proj_weight_quantizers.108 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.down_proj_weight_quantizers.109 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.down_proj_weight_quantizers.110 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.59e-02 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.down_proj_weight_quantizers.111 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.23e-01 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.down_proj_weight_quantizers.112 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.down_proj_weight_quantizers.113 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.28e-01 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.down_proj_weight_quantizers.114 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.54e-02 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.down_proj_weight_quantizers.115 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.41e-01 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.down_proj_weight_quantizers.116 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.50e-02 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.down_proj_weight_quantizers.117 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.down_proj_weight_quantizers.118 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.down_proj_weight_quantizers.119 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.down_proj_weight_quantizers.120 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.27e-01 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.down_proj_weight_quantizers.121 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.39e-01 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.down_proj_weight_quantizers.122 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.down_proj_weight_quantizers.123 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.down_proj_weight_quantizers.124 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.59e-02 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.down_proj_weight_quantizers.125 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.down_proj_weight_quantizers.126 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.18e-01 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.down_proj_weight_quantizers.127 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.25e-01 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.down_proj_weight_quantizers.128 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.40e-02 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.down_proj_weight_quantizers.129 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.down_proj_weight_quantizers.130 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.41e-01 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.down_proj_weight_quantizers.131 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.down_proj_weight_quantizers.132 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.down_proj_weight_quantizers.133 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.down_proj_weight_quantizers.134 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.down_proj_weight_quantizers.135 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.down_proj_weight_quantizers.136 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.19e-01 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.down_proj_weight_quantizers.137 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.down_proj_weight_quantizers.138 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.down_proj_weight_quantizers.139 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.down_proj_weight_quantizers.140 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.down_proj_weight_quantizers.141 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.down_proj_weight_quantizers.142 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.down_proj_weight_quantizers.143 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.down_proj_weight_quantizers.144 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.50e-02 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.down_proj_weight_quantizers.145 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.down_proj_weight_quantizers.146 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.97e-01 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.down_proj_weight_quantizers.147 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.down_proj_weight_quantizers.148 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.37e-01 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.down_proj_weight_quantizers.149 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.down_proj_weight_quantizers.150 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.40e-02 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.down_proj_weight_quantizers.151 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.down_proj_weight_quantizers.152 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.down_proj_weight_quantizers.153 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.down_proj_weight_quantizers.154 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.30e-02 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.down_proj_weight_quantizers.155 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.down_proj_weight_quantizers.156 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.20e-01 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.down_proj_weight_quantizers.157 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.down_proj_weight_quantizers.158 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.59e-02 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.down_proj_weight_quantizers.159 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.down_proj_weight_quantizers.160 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.64e-02 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.down_proj_weight_quantizers.161 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.down_proj_weight_quantizers.162 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.40e-02 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.down_proj_weight_quantizers.163 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.down_proj_weight_quantizers.164 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.down_proj_weight_quantizers.165 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.down_proj_weight_quantizers.166 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.54e-02 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.down_proj_weight_quantizers.167 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.down_proj_weight_quantizers.168 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.down_proj_weight_quantizers.169 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.down_proj_weight_quantizers.170 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.down_proj_weight_quantizers.171 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.down_proj_weight_quantizers.172 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.down_proj_weight_quantizers.173 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.59e-02 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.down_proj_weight_quantizers.174 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.down_proj_weight_quantizers.175 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.77e-01 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.down_proj_weight_quantizers.176 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.down_proj_weight_quantizers.177 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.down_proj_weight_quantizers.178 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.down_proj_weight_quantizers.179 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.down_proj_weight_quantizers.180 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.down_proj_weight_quantizers.181 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.down_proj_weight_quantizers.182 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.down_proj_weight_quantizers.183 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.down_proj_weight_quantizers.184 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.down_proj_weight_quantizers.185 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.down_proj_weight_quantizers.186 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.down_proj_weight_quantizers.187 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.down_proj_weight_quantizers.188 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.down_proj_weight_quantizers.189 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.59e-02 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.down_proj_weight_quantizers.190 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.54e-02 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.down_proj_weight_quantizers.191 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.down_proj_weight_quantizers.192 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.down_proj_weight_quantizers.193 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.down_proj_weight_quantizers.194 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.down_proj_weight_quantizers.195 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.down_proj_weight_quantizers.196 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.40e-02 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.down_proj_weight_quantizers.197 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.42e-01 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.down_proj_weight_quantizers.198 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.45e-02 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.down_proj_weight_quantizers.199 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.down_proj_weight_quantizers.200 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.down_proj_weight_quantizers.201 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.down_proj_weight_quantizers.202 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.down_proj_weight_quantizers.203 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.down_proj_weight_quantizers.204 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.down_proj_weight_quantizers.205 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.64e-01 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.down_proj_weight_quantizers.206 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.down_proj_weight_quantizers.207 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.down_proj_weight_quantizers.208 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.down_proj_weight_quantizers.209 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.down_proj_weight_quantizers.210 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.59e-02 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.down_proj_weight_quantizers.211 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.down_proj_weight_quantizers.212 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.down_proj_weight_quantizers.213 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.down_proj_weight_quantizers.214 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.20e-01 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.down_proj_weight_quantizers.215 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.down_proj_weight_quantizers.216 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.82e-01 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.down_proj_weight_quantizers.217 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.down_proj_weight_quantizers.218 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.45e-02 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.down_proj_weight_quantizers.219 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.down_proj_weight_quantizers.220 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.down_proj_weight_quantizers.221 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.down_proj_weight_quantizers.222 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.down_proj_weight_quantizers.223 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.down_proj_weight_quantizers.224 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.down_proj_weight_quantizers.225 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.down_proj_weight_quantizers.226 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.64e-02 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.down_proj_weight_quantizers.227 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.down_proj_weight_quantizers.228 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.down_proj_weight_quantizers.229 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.down_proj_weight_quantizers.230 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.down_proj_weight_quantizers.231 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.down_proj_weight_quantizers.232 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.down_proj_weight_quantizers.233 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.down_proj_weight_quantizers.234 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.down_proj_weight_quantizers.235 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.down_proj_weight_quantizers.236 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.50e-02 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.down_proj_weight_quantizers.237 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.down_proj_weight_quantizers.238 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.down_proj_weight_quantizers.239 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.down_proj_weight_quantizers.240 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.down_proj_weight_quantizers.241 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.down_proj_weight_quantizers.242 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.down_proj_weight_quantizers.243 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.down_proj_weight_quantizers.244 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.53e-01 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.down_proj_weight_quantizers.245 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.down_proj_weight_quantizers.246 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.down_proj_weight_quantizers.247 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.down_proj_weight_quantizers.248 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.32e-01 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.down_proj_weight_quantizers.249 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.49e-01 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.down_proj_weight_quantizers.250 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.down_proj_weight_quantizers.251 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.45e-02 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.down_proj_weight_quantizers.252 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.down_proj_weight_quantizers.253 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.45e-02 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.down_proj_weight_quantizers.254 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.down_proj_weight_quantizers.255 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.21.mlp.shared_experts.gate_proj.input_quantizer TensorQuantizer(disabled) +model.layers.21.mlp.shared_experts.gate_proj.output_quantizer TensorQuantizer(disabled) +model.layers.21.mlp.shared_experts.gate_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.21.mlp.shared_experts.up_proj.input_quantizer TensorQuantizer(disabled) +model.layers.21.mlp.shared_experts.up_proj.output_quantizer TensorQuantizer(disabled) +model.layers.21.mlp.shared_experts.up_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.21.mlp.shared_experts.down_proj.input_quantizer TensorQuantizer(disabled) +model.layers.21.mlp.shared_experts.down_proj.output_quantizer TensorQuantizer(disabled) +model.layers.21.mlp.shared_experts.down_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.22.self_attn.q_a_proj.input_quantizer TensorQuantizer(disabled) +model.layers.22.self_attn.q_a_proj.output_quantizer TensorQuantizer(disabled) +model.layers.22.self_attn.q_a_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.22.self_attn.q_b_proj.input_quantizer TensorQuantizer(disabled) +model.layers.22.self_attn.q_b_proj.output_quantizer TensorQuantizer(disabled) +model.layers.22.self_attn.q_b_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.22.self_attn.kv_a_proj_with_mqa.input_quantizer TensorQuantizer(disabled) +model.layers.22.self_attn.kv_a_proj_with_mqa.output_quantizer TensorQuantizer(disabled) +model.layers.22.self_attn.kv_a_proj_with_mqa.weight_quantizer TensorQuantizer(disabled) +model.layers.22.self_attn.kv_b_proj.input_quantizer TensorQuantizer(disabled) +model.layers.22.self_attn.kv_b_proj.output_quantizer TensorQuantizer(disabled) +model.layers.22.self_attn.kv_b_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.22.self_attn.o_proj.input_quantizer TensorQuantizer(disabled) +model.layers.22.self_attn.o_proj.output_quantizer TensorQuantizer(disabled) +model.layers.22.self_attn.o_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.22.self_attn.indexer.wq_b.input_quantizer TensorQuantizer(disabled) +model.layers.22.self_attn.indexer.wq_b.output_quantizer TensorQuantizer(disabled) +model.layers.22.self_attn.indexer.wq_b.weight_quantizer TensorQuantizer(disabled) +model.layers.22.self_attn.indexer.wk.input_quantizer TensorQuantizer(disabled) +model.layers.22.self_attn.indexer.wk.output_quantizer TensorQuantizer(disabled) +model.layers.22.self_attn.indexer.wk.weight_quantizer TensorQuantizer(disabled) +model.layers.22.self_attn.indexer.k_norm.input_quantizer TensorQuantizer(disabled) +model.layers.22.self_attn.indexer.k_norm.output_quantizer TensorQuantizer(disabled) +model.layers.22.self_attn.indexer.weights_proj.input_quantizer TensorQuantizer(disabled) +model.layers.22.self_attn.indexer.weights_proj.output_quantizer TensorQuantizer(disabled) +model.layers.22.self_attn.indexer.weights_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.22.self_attn.q_bmm_quantizer TensorQuantizer(disabled) +model.layers.22.self_attn.k_bmm_quantizer TensorQuantizer((4, 3) bit fake per-tensor amax=4.48e+02(const) calibrator=MaxCalibrator quant) +model.layers.22.self_attn.v_bmm_quantizer TensorQuantizer((4, 3) bit fake per-tensor amax=4.48e+02(const) calibrator=MaxCalibrator quant) +model.layers.22.self_attn.p_bmm_quantizer TensorQuantizer(disabled) +model.layers.22.mlp.experts.gate_up_proj_input_quantizer TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.48e-01 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.gate_up_proj_weight_quantizers.0 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.gate_up_proj_weight_quantizers.1 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.gate_up_proj_weight_quantizers.2 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.gate_up_proj_weight_quantizers.3 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.gate_up_proj_weight_quantizers.4 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.gate_up_proj_weight_quantizers.5 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.gate_up_proj_weight_quantizers.6 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.gate_up_proj_weight_quantizers.7 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.gate_up_proj_weight_quantizers.8 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.gate_up_proj_weight_quantizers.9 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.gate_up_proj_weight_quantizers.10 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.43e-01 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.gate_up_proj_weight_quantizers.11 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.22e-01 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.gate_up_proj_weight_quantizers.12 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.gate_up_proj_weight_quantizers.13 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.19e-01 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.gate_up_proj_weight_quantizers.14 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.gate_up_proj_weight_quantizers.15 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.22e-01 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.gate_up_proj_weight_quantizers.16 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.gate_up_proj_weight_quantizers.17 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.11e-01 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.gate_up_proj_weight_quantizers.18 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.gate_up_proj_weight_quantizers.19 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.gate_up_proj_weight_quantizers.20 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.43e-01 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.gate_up_proj_weight_quantizers.21 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.gate_up_proj_weight_quantizers.22 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.gate_up_proj_weight_quantizers.23 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.gate_up_proj_weight_quantizers.24 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.27e-01 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.gate_up_proj_weight_quantizers.25 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.gate_up_proj_weight_quantizers.26 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.gate_up_proj_weight_quantizers.27 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.gate_up_proj_weight_quantizers.28 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.gate_up_proj_weight_quantizers.29 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.gate_up_proj_weight_quantizers.30 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.gate_up_proj_weight_quantizers.31 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.20e-01 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.gate_up_proj_weight_quantizers.32 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.gate_up_proj_weight_quantizers.33 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.14e-01 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.gate_up_proj_weight_quantizers.34 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.gate_up_proj_weight_quantizers.35 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.gate_up_proj_weight_quantizers.36 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.gate_up_proj_weight_quantizers.37 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.gate_up_proj_weight_quantizers.38 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.18e-01 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.gate_up_proj_weight_quantizers.39 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.gate_up_proj_weight_quantizers.40 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.gate_up_proj_weight_quantizers.41 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.25e-01 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.gate_up_proj_weight_quantizers.42 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.gate_up_proj_weight_quantizers.43 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.23e-01 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.gate_up_proj_weight_quantizers.44 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.gate_up_proj_weight_quantizers.45 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.gate_up_proj_weight_quantizers.46 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.gate_up_proj_weight_quantizers.47 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.gate_up_proj_weight_quantizers.48 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.13e-01 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.gate_up_proj_weight_quantizers.49 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.gate_up_proj_weight_quantizers.50 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.gate_up_proj_weight_quantizers.51 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.gate_up_proj_weight_quantizers.52 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.gate_up_proj_weight_quantizers.53 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.gate_up_proj_weight_quantizers.54 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.51e-01 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.gate_up_proj_weight_quantizers.55 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.gate_up_proj_weight_quantizers.56 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.46e-01 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.gate_up_proj_weight_quantizers.57 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.25e-01 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.gate_up_proj_weight_quantizers.58 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.gate_up_proj_weight_quantizers.59 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.gate_up_proj_weight_quantizers.60 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.gate_up_proj_weight_quantizers.61 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.gate_up_proj_weight_quantizers.62 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.gate_up_proj_weight_quantizers.63 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.gate_up_proj_weight_quantizers.64 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.33e-01 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.gate_up_proj_weight_quantizers.65 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.59e-01 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.gate_up_proj_weight_quantizers.66 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.gate_up_proj_weight_quantizers.67 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.gate_up_proj_weight_quantizers.68 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.gate_up_proj_weight_quantizers.69 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.gate_up_proj_weight_quantizers.70 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.gate_up_proj_weight_quantizers.71 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.91e-01 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.gate_up_proj_weight_quantizers.72 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.gate_up_proj_weight_quantizers.73 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.02e-01 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.gate_up_proj_weight_quantizers.74 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.gate_up_proj_weight_quantizers.75 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.gate_up_proj_weight_quantizers.76 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.gate_up_proj_weight_quantizers.77 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.03e-01 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.gate_up_proj_weight_quantizers.78 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.21e-01 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.gate_up_proj_weight_quantizers.79 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.52e-01 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.gate_up_proj_weight_quantizers.80 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.gate_up_proj_weight_quantizers.81 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.60e-01 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.gate_up_proj_weight_quantizers.82 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.gate_up_proj_weight_quantizers.83 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.gate_up_proj_weight_quantizers.84 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.gate_up_proj_weight_quantizers.85 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.gate_up_proj_weight_quantizers.86 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.gate_up_proj_weight_quantizers.87 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.gate_up_proj_weight_quantizers.88 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.35e-01 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.gate_up_proj_weight_quantizers.89 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.gate_up_proj_weight_quantizers.90 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.18e-01 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.gate_up_proj_weight_quantizers.91 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.17e-01 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.gate_up_proj_weight_quantizers.92 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.26e-01 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.gate_up_proj_weight_quantizers.93 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.gate_up_proj_weight_quantizers.94 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.gate_up_proj_weight_quantizers.95 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.gate_up_proj_weight_quantizers.96 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.gate_up_proj_weight_quantizers.97 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.29e-01 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.gate_up_proj_weight_quantizers.98 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.85e-01 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.gate_up_proj_weight_quantizers.99 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.gate_up_proj_weight_quantizers.100 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.gate_up_proj_weight_quantizers.101 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.25e-01 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.gate_up_proj_weight_quantizers.102 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.gate_up_proj_weight_quantizers.103 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.gate_up_proj_weight_quantizers.104 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.gate_up_proj_weight_quantizers.105 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.gate_up_proj_weight_quantizers.106 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.gate_up_proj_weight_quantizers.107 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.gate_up_proj_weight_quantizers.108 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.gate_up_proj_weight_quantizers.109 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.gate_up_proj_weight_quantizers.110 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.gate_up_proj_weight_quantizers.111 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.gate_up_proj_weight_quantizers.112 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.gate_up_proj_weight_quantizers.113 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.gate_up_proj_weight_quantizers.114 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.gate_up_proj_weight_quantizers.115 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.24e-01 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.gate_up_proj_weight_quantizers.116 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.gate_up_proj_weight_quantizers.117 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.gate_up_proj_weight_quantizers.118 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.gate_up_proj_weight_quantizers.119 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.gate_up_proj_weight_quantizers.120 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.gate_up_proj_weight_quantizers.121 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.38e-01 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.gate_up_proj_weight_quantizers.122 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.57e-01 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.gate_up_proj_weight_quantizers.123 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.gate_up_proj_weight_quantizers.124 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.37e-01 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.gate_up_proj_weight_quantizers.125 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.gate_up_proj_weight_quantizers.126 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.58e-01 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.gate_up_proj_weight_quantizers.127 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.gate_up_proj_weight_quantizers.128 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.45e-01 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.gate_up_proj_weight_quantizers.129 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.32e-01 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.gate_up_proj_weight_quantizers.130 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.gate_up_proj_weight_quantizers.131 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.gate_up_proj_weight_quantizers.132 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.99e-01 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.gate_up_proj_weight_quantizers.133 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.gate_up_proj_weight_quantizers.134 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.gate_up_proj_weight_quantizers.135 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.gate_up_proj_weight_quantizers.136 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.71e-01 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.gate_up_proj_weight_quantizers.137 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.gate_up_proj_weight_quantizers.138 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.gate_up_proj_weight_quantizers.139 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.15e-01 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.gate_up_proj_weight_quantizers.140 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.gate_up_proj_weight_quantizers.141 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.93e-01 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.gate_up_proj_weight_quantizers.142 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.gate_up_proj_weight_quantizers.143 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.55e-01 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.gate_up_proj_weight_quantizers.144 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.gate_up_proj_weight_quantizers.145 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.46e-01 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.gate_up_proj_weight_quantizers.146 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.gate_up_proj_weight_quantizers.147 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.45e-01 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.gate_up_proj_weight_quantizers.148 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.gate_up_proj_weight_quantizers.149 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.15e-01 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.gate_up_proj_weight_quantizers.150 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.gate_up_proj_weight_quantizers.151 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.gate_up_proj_weight_quantizers.152 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.gate_up_proj_weight_quantizers.153 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.gate_up_proj_weight_quantizers.154 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.83e-01 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.gate_up_proj_weight_quantizers.155 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.gate_up_proj_weight_quantizers.156 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.20e-01 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.gate_up_proj_weight_quantizers.157 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.gate_up_proj_weight_quantizers.158 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.gate_up_proj_weight_quantizers.159 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.gate_up_proj_weight_quantizers.160 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.gate_up_proj_weight_quantizers.161 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.gate_up_proj_weight_quantizers.162 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.gate_up_proj_weight_quantizers.163 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.gate_up_proj_weight_quantizers.164 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.gate_up_proj_weight_quantizers.165 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.gate_up_proj_weight_quantizers.166 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.gate_up_proj_weight_quantizers.167 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.gate_up_proj_weight_quantizers.168 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.gate_up_proj_weight_quantizers.169 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.gate_up_proj_weight_quantizers.170 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.gate_up_proj_weight_quantizers.171 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.gate_up_proj_weight_quantizers.172 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.gate_up_proj_weight_quantizers.173 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.gate_up_proj_weight_quantizers.174 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.gate_up_proj_weight_quantizers.175 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.09e-01 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.gate_up_proj_weight_quantizers.176 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.gate_up_proj_weight_quantizers.177 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.gate_up_proj_weight_quantizers.178 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.gate_up_proj_weight_quantizers.179 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.gate_up_proj_weight_quantizers.180 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.gate_up_proj_weight_quantizers.181 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=4.73e-01 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.gate_up_proj_weight_quantizers.182 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.gate_up_proj_weight_quantizers.183 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.gate_up_proj_weight_quantizers.184 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.gate_up_proj_weight_quantizers.185 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.gate_up_proj_weight_quantizers.186 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.23e-01 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.gate_up_proj_weight_quantizers.187 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.gate_up_proj_weight_quantizers.188 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.gate_up_proj_weight_quantizers.189 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.19e-01 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.gate_up_proj_weight_quantizers.190 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.gate_up_proj_weight_quantizers.191 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.gate_up_proj_weight_quantizers.192 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.57e-01 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.gate_up_proj_weight_quantizers.193 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.gate_up_proj_weight_quantizers.194 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.24e-01 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.gate_up_proj_weight_quantizers.195 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.gate_up_proj_weight_quantizers.196 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.gate_up_proj_weight_quantizers.197 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.39e-01 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.gate_up_proj_weight_quantizers.198 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.73e-01 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.gate_up_proj_weight_quantizers.199 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.gate_up_proj_weight_quantizers.200 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.gate_up_proj_weight_quantizers.201 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.gate_up_proj_weight_quantizers.202 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.73e-01 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.gate_up_proj_weight_quantizers.203 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.gate_up_proj_weight_quantizers.204 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.59e-02 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.gate_up_proj_weight_quantizers.205 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.gate_up_proj_weight_quantizers.206 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.gate_up_proj_weight_quantizers.207 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.59e-01 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.gate_up_proj_weight_quantizers.208 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.gate_up_proj_weight_quantizers.209 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.gate_up_proj_weight_quantizers.210 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.gate_up_proj_weight_quantizers.211 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.78e-01 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.gate_up_proj_weight_quantizers.212 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.48e-01 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.gate_up_proj_weight_quantizers.213 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.gate_up_proj_weight_quantizers.214 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.gate_up_proj_weight_quantizers.215 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.gate_up_proj_weight_quantizers.216 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.gate_up_proj_weight_quantizers.217 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.gate_up_proj_weight_quantizers.218 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.gate_up_proj_weight_quantizers.219 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.gate_up_proj_weight_quantizers.220 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.gate_up_proj_weight_quantizers.221 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.gate_up_proj_weight_quantizers.222 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.38e-01 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.gate_up_proj_weight_quantizers.223 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.gate_up_proj_weight_quantizers.224 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.64e-02 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.gate_up_proj_weight_quantizers.225 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.gate_up_proj_weight_quantizers.226 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.gate_up_proj_weight_quantizers.227 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.gate_up_proj_weight_quantizers.228 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.82e-01 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.gate_up_proj_weight_quantizers.229 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.75e-01 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.gate_up_proj_weight_quantizers.230 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.gate_up_proj_weight_quantizers.231 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.gate_up_proj_weight_quantizers.232 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.gate_up_proj_weight_quantizers.233 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.38e-01 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.gate_up_proj_weight_quantizers.234 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.gate_up_proj_weight_quantizers.235 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.gate_up_proj_weight_quantizers.236 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.gate_up_proj_weight_quantizers.237 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.gate_up_proj_weight_quantizers.238 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.41e-01 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.gate_up_proj_weight_quantizers.239 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.gate_up_proj_weight_quantizers.240 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.gate_up_proj_weight_quantizers.241 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.gate_up_proj_weight_quantizers.242 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.gate_up_proj_weight_quantizers.243 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.gate_up_proj_weight_quantizers.244 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=4.08e-01 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.gate_up_proj_weight_quantizers.245 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.gate_up_proj_weight_quantizers.246 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.gate_up_proj_weight_quantizers.247 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.gate_up_proj_weight_quantizers.248 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.gate_up_proj_weight_quantizers.249 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.21e-01 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.gate_up_proj_weight_quantizers.250 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.gate_up_proj_weight_quantizers.251 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.gate_up_proj_weight_quantizers.252 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.gate_up_proj_weight_quantizers.253 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.00e-01 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.gate_up_proj_weight_quantizers.254 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.74e-01 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.gate_up_proj_weight_quantizers.255 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.down_proj_input_quantizer TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=5.62e+00 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.down_proj_weight_quantizers.0 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.down_proj_weight_quantizers.1 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.down_proj_weight_quantizers.2 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.down_proj_weight_quantizers.3 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.down_proj_weight_quantizers.4 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.down_proj_weight_quantizers.5 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.down_proj_weight_quantizers.6 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.down_proj_weight_quantizers.7 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.down_proj_weight_quantizers.8 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.down_proj_weight_quantizers.9 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.down_proj_weight_quantizers.10 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.43e-01 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.down_proj_weight_quantizers.11 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.down_proj_weight_quantizers.12 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.64e-02 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.down_proj_weight_quantizers.13 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.down_proj_weight_quantizers.14 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.40e-02 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.down_proj_weight_quantizers.15 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.53e-01 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.down_proj_weight_quantizers.16 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.down_proj_weight_quantizers.17 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.84e-01 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.down_proj_weight_quantizers.18 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.15e-01 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.down_proj_weight_quantizers.19 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.59e-02 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.down_proj_weight_quantizers.20 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.down_proj_weight_quantizers.21 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.down_proj_weight_quantizers.22 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.down_proj_weight_quantizers.23 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.down_proj_weight_quantizers.24 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.down_proj_weight_quantizers.25 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.down_proj_weight_quantizers.26 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.48e-01 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.down_proj_weight_quantizers.27 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.down_proj_weight_quantizers.28 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.down_proj_weight_quantizers.29 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.down_proj_weight_quantizers.30 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.down_proj_weight_quantizers.31 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.16e-01 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.down_proj_weight_quantizers.32 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.down_proj_weight_quantizers.33 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.down_proj_weight_quantizers.34 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.down_proj_weight_quantizers.35 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.40e-02 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.down_proj_weight_quantizers.36 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.down_proj_weight_quantizers.37 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.down_proj_weight_quantizers.38 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.96e-01 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.down_proj_weight_quantizers.39 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.down_proj_weight_quantizers.40 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.45e-02 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.down_proj_weight_quantizers.41 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.down_proj_weight_quantizers.42 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.down_proj_weight_quantizers.43 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.down_proj_weight_quantizers.44 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.down_proj_weight_quantizers.45 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.down_proj_weight_quantizers.46 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.51e-01 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.down_proj_weight_quantizers.47 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.24e-01 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.down_proj_weight_quantizers.48 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.down_proj_weight_quantizers.49 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.down_proj_weight_quantizers.50 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.down_proj_weight_quantizers.51 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.64e-02 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.down_proj_weight_quantizers.52 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.64e-02 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.down_proj_weight_quantizers.53 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.down_proj_weight_quantizers.54 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.33e-01 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.down_proj_weight_quantizers.55 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.down_proj_weight_quantizers.56 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.45e-01 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.down_proj_weight_quantizers.57 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.down_proj_weight_quantizers.58 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.down_proj_weight_quantizers.59 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.59e-02 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.down_proj_weight_quantizers.60 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.25e-02 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.down_proj_weight_quantizers.61 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.29e-01 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.down_proj_weight_quantizers.62 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.down_proj_weight_quantizers.63 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.down_proj_weight_quantizers.64 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.30e-01 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.down_proj_weight_quantizers.65 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.20e-02 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.down_proj_weight_quantizers.66 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.down_proj_weight_quantizers.67 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.down_proj_weight_quantizers.68 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.down_proj_weight_quantizers.69 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.down_proj_weight_quantizers.70 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.down_proj_weight_quantizers.71 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.48e-01 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.down_proj_weight_quantizers.72 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.down_proj_weight_quantizers.73 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.57e-01 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.down_proj_weight_quantizers.74 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.down_proj_weight_quantizers.75 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.18e-01 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.down_proj_weight_quantizers.76 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.down_proj_weight_quantizers.77 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.down_proj_weight_quantizers.78 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.29e-01 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.down_proj_weight_quantizers.79 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.down_proj_weight_quantizers.80 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.down_proj_weight_quantizers.81 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.56e-01 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.down_proj_weight_quantizers.82 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.down_proj_weight_quantizers.83 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.64e-02 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.down_proj_weight_quantizers.84 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.down_proj_weight_quantizers.85 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.down_proj_weight_quantizers.86 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.down_proj_weight_quantizers.87 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.down_proj_weight_quantizers.88 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.60e-01 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.down_proj_weight_quantizers.89 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.down_proj_weight_quantizers.90 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.58e-01 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.down_proj_weight_quantizers.91 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.down_proj_weight_quantizers.92 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.46e-01 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.down_proj_weight_quantizers.93 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.down_proj_weight_quantizers.94 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.down_proj_weight_quantizers.95 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.down_proj_weight_quantizers.96 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.down_proj_weight_quantizers.97 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.42e-01 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.down_proj_weight_quantizers.98 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.down_proj_weight_quantizers.99 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.down_proj_weight_quantizers.100 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.45e-02 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.down_proj_weight_quantizers.101 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.43e-01 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.down_proj_weight_quantizers.102 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.down_proj_weight_quantizers.103 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.down_proj_weight_quantizers.104 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.down_proj_weight_quantizers.105 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.64e-02 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.down_proj_weight_quantizers.106 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.down_proj_weight_quantizers.107 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.down_proj_weight_quantizers.108 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.down_proj_weight_quantizers.109 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.down_proj_weight_quantizers.110 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.down_proj_weight_quantizers.111 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.down_proj_weight_quantizers.112 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.down_proj_weight_quantizers.113 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.down_proj_weight_quantizers.114 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.down_proj_weight_quantizers.115 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.27e-01 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.down_proj_weight_quantizers.116 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.down_proj_weight_quantizers.117 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.64e-02 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.down_proj_weight_quantizers.118 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.down_proj_weight_quantizers.119 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.down_proj_weight_quantizers.120 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.down_proj_weight_quantizers.121 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.53e-01 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.down_proj_weight_quantizers.122 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.down_proj_weight_quantizers.123 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.down_proj_weight_quantizers.124 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.08e-01 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.down_proj_weight_quantizers.125 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.64e-02 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.down_proj_weight_quantizers.126 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.down_proj_weight_quantizers.127 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.down_proj_weight_quantizers.128 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.02e-01 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.down_proj_weight_quantizers.129 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.down_proj_weight_quantizers.130 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.down_proj_weight_quantizers.131 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.down_proj_weight_quantizers.132 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.down_proj_weight_quantizers.133 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.64e-02 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.down_proj_weight_quantizers.134 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.down_proj_weight_quantizers.135 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.64e-02 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.down_proj_weight_quantizers.136 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.down_proj_weight_quantizers.137 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.down_proj_weight_quantizers.138 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.down_proj_weight_quantizers.139 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.down_proj_weight_quantizers.140 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.down_proj_weight_quantizers.141 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.54e-01 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.down_proj_weight_quantizers.142 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.down_proj_weight_quantizers.143 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.down_proj_weight_quantizers.144 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.24e-01 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.down_proj_weight_quantizers.145 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.14e-01 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.down_proj_weight_quantizers.146 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.down_proj_weight_quantizers.147 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.down_proj_weight_quantizers.148 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.26e-01 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.down_proj_weight_quantizers.149 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.down_proj_weight_quantizers.150 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.down_proj_weight_quantizers.151 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.down_proj_weight_quantizers.152 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.down_proj_weight_quantizers.153 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.down_proj_weight_quantizers.154 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.52e-01 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.down_proj_weight_quantizers.155 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.down_proj_weight_quantizers.156 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.26e-01 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.down_proj_weight_quantizers.157 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.down_proj_weight_quantizers.158 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.13e-01 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.down_proj_weight_quantizers.159 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.45e-02 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.down_proj_weight_quantizers.160 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.down_proj_weight_quantizers.161 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.down_proj_weight_quantizers.162 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.down_proj_weight_quantizers.163 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.down_proj_weight_quantizers.164 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.down_proj_weight_quantizers.165 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.down_proj_weight_quantizers.166 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.down_proj_weight_quantizers.167 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.down_proj_weight_quantizers.168 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.down_proj_weight_quantizers.169 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.down_proj_weight_quantizers.170 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.64e-02 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.down_proj_weight_quantizers.171 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.71e-01 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.down_proj_weight_quantizers.172 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.down_proj_weight_quantizers.173 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.59e-02 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.down_proj_weight_quantizers.174 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.down_proj_weight_quantizers.175 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.50e-01 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.down_proj_weight_quantizers.176 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.down_proj_weight_quantizers.177 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.50e-02 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.down_proj_weight_quantizers.178 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.down_proj_weight_quantizers.179 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.down_proj_weight_quantizers.180 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.down_proj_weight_quantizers.181 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.99e-01 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.down_proj_weight_quantizers.182 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.down_proj_weight_quantizers.183 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.down_proj_weight_quantizers.184 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.down_proj_weight_quantizers.185 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.down_proj_weight_quantizers.186 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.21e-01 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.down_proj_weight_quantizers.187 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.down_proj_weight_quantizers.188 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.down_proj_weight_quantizers.189 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.74e-01 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.down_proj_weight_quantizers.190 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.down_proj_weight_quantizers.191 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.down_proj_weight_quantizers.192 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.down_proj_weight_quantizers.193 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.down_proj_weight_quantizers.194 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.44e-01 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.down_proj_weight_quantizers.195 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.down_proj_weight_quantizers.196 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.down_proj_weight_quantizers.197 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.57e-01 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.down_proj_weight_quantizers.198 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.down_proj_weight_quantizers.199 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.50e-02 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.down_proj_weight_quantizers.200 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.down_proj_weight_quantizers.201 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.down_proj_weight_quantizers.202 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.down_proj_weight_quantizers.203 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.54e-02 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.down_proj_weight_quantizers.204 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.down_proj_weight_quantizers.205 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.down_proj_weight_quantizers.206 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.down_proj_weight_quantizers.207 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.15e-01 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.down_proj_weight_quantizers.208 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.down_proj_weight_quantizers.209 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.down_proj_weight_quantizers.210 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.down_proj_weight_quantizers.211 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.80e-01 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.down_proj_weight_quantizers.212 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.down_proj_weight_quantizers.213 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.down_proj_weight_quantizers.214 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.down_proj_weight_quantizers.215 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.59e-02 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.down_proj_weight_quantizers.216 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.down_proj_weight_quantizers.217 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.down_proj_weight_quantizers.218 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.down_proj_weight_quantizers.219 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.down_proj_weight_quantizers.220 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.88e-01 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.down_proj_weight_quantizers.221 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.down_proj_weight_quantizers.222 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.44e-01 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.down_proj_weight_quantizers.223 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.down_proj_weight_quantizers.224 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.down_proj_weight_quantizers.225 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.down_proj_weight_quantizers.226 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.down_proj_weight_quantizers.227 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.down_proj_weight_quantizers.228 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.20e-01 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.down_proj_weight_quantizers.229 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.43e-01 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.down_proj_weight_quantizers.230 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.down_proj_weight_quantizers.231 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.down_proj_weight_quantizers.232 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.64e-02 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.down_proj_weight_quantizers.233 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.down_proj_weight_quantizers.234 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.down_proj_weight_quantizers.235 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.down_proj_weight_quantizers.236 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.down_proj_weight_quantizers.237 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.down_proj_weight_quantizers.238 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.down_proj_weight_quantizers.239 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.down_proj_weight_quantizers.240 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.down_proj_weight_quantizers.241 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.down_proj_weight_quantizers.242 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.18e-01 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.down_proj_weight_quantizers.243 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.97e-01 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.down_proj_weight_quantizers.244 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.77e-01 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.down_proj_weight_quantizers.245 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.down_proj_weight_quantizers.246 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.down_proj_weight_quantizers.247 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.down_proj_weight_quantizers.248 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.down_proj_weight_quantizers.249 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.88e-01 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.down_proj_weight_quantizers.250 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.down_proj_weight_quantizers.251 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.down_proj_weight_quantizers.252 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.down_proj_weight_quantizers.253 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.49e-01 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.down_proj_weight_quantizers.254 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.91e-01 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.down_proj_weight_quantizers.255 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.22.mlp.shared_experts.gate_proj.input_quantizer TensorQuantizer(disabled) +model.layers.22.mlp.shared_experts.gate_proj.output_quantizer TensorQuantizer(disabled) +model.layers.22.mlp.shared_experts.gate_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.22.mlp.shared_experts.up_proj.input_quantizer TensorQuantizer(disabled) +model.layers.22.mlp.shared_experts.up_proj.output_quantizer TensorQuantizer(disabled) +model.layers.22.mlp.shared_experts.up_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.22.mlp.shared_experts.down_proj.input_quantizer TensorQuantizer(disabled) +model.layers.22.mlp.shared_experts.down_proj.output_quantizer TensorQuantizer(disabled) +model.layers.22.mlp.shared_experts.down_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.23.self_attn.q_a_proj.input_quantizer TensorQuantizer(disabled) +model.layers.23.self_attn.q_a_proj.output_quantizer TensorQuantizer(disabled) +model.layers.23.self_attn.q_a_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.23.self_attn.q_b_proj.input_quantizer TensorQuantizer(disabled) +model.layers.23.self_attn.q_b_proj.output_quantizer TensorQuantizer(disabled) +model.layers.23.self_attn.q_b_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.23.self_attn.kv_a_proj_with_mqa.input_quantizer TensorQuantizer(disabled) +model.layers.23.self_attn.kv_a_proj_with_mqa.output_quantizer TensorQuantizer(disabled) +model.layers.23.self_attn.kv_a_proj_with_mqa.weight_quantizer TensorQuantizer(disabled) +model.layers.23.self_attn.kv_b_proj.input_quantizer TensorQuantizer(disabled) +model.layers.23.self_attn.kv_b_proj.output_quantizer TensorQuantizer(disabled) +model.layers.23.self_attn.kv_b_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.23.self_attn.o_proj.input_quantizer TensorQuantizer(disabled) +model.layers.23.self_attn.o_proj.output_quantizer TensorQuantizer(disabled) +model.layers.23.self_attn.o_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.23.self_attn.q_bmm_quantizer TensorQuantizer(disabled) +model.layers.23.self_attn.k_bmm_quantizer TensorQuantizer((4, 3) bit fake per-tensor amax=4.48e+02(const) calibrator=MaxCalibrator quant) +model.layers.23.self_attn.v_bmm_quantizer TensorQuantizer((4, 3) bit fake per-tensor amax=4.48e+02(const) calibrator=MaxCalibrator quant) +model.layers.23.self_attn.p_bmm_quantizer TensorQuantizer(disabled) +model.layers.23.mlp.experts.gate_up_proj_input_quantizer TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.63e-01 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.gate_up_proj_weight_quantizers.0 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.gate_up_proj_weight_quantizers.1 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.gate_up_proj_weight_quantizers.2 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.64e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.gate_up_proj_weight_quantizers.3 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.gate_up_proj_weight_quantizers.4 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.gate_up_proj_weight_quantizers.5 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.gate_up_proj_weight_quantizers.6 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.gate_up_proj_weight_quantizers.7 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.gate_up_proj_weight_quantizers.8 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.gate_up_proj_weight_quantizers.9 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.gate_up_proj_weight_quantizers.10 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.gate_up_proj_weight_quantizers.11 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.gate_up_proj_weight_quantizers.12 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.23e-01 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.gate_up_proj_weight_quantizers.13 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.gate_up_proj_weight_quantizers.14 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.gate_up_proj_weight_quantizers.15 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.gate_up_proj_weight_quantizers.16 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.gate_up_proj_weight_quantizers.17 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.gate_up_proj_weight_quantizers.18 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.gate_up_proj_weight_quantizers.19 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.gate_up_proj_weight_quantizers.20 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.gate_up_proj_weight_quantizers.21 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.gate_up_proj_weight_quantizers.22 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.gate_up_proj_weight_quantizers.23 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.13e-01 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.gate_up_proj_weight_quantizers.24 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.gate_up_proj_weight_quantizers.25 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.gate_up_proj_weight_quantizers.26 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.77e-01 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.gate_up_proj_weight_quantizers.27 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.33e-01 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.gate_up_proj_weight_quantizers.28 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.gate_up_proj_weight_quantizers.29 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.70e-01 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.gate_up_proj_weight_quantizers.30 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.gate_up_proj_weight_quantizers.31 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.19e-01 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.gate_up_proj_weight_quantizers.32 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.gate_up_proj_weight_quantizers.33 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.gate_up_proj_weight_quantizers.34 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.gate_up_proj_weight_quantizers.35 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.gate_up_proj_weight_quantizers.36 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.gate_up_proj_weight_quantizers.37 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.gate_up_proj_weight_quantizers.38 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.gate_up_proj_weight_quantizers.39 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.gate_up_proj_weight_quantizers.40 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.gate_up_proj_weight_quantizers.41 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.gate_up_proj_weight_quantizers.42 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.gate_up_proj_weight_quantizers.43 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.gate_up_proj_weight_quantizers.44 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.gate_up_proj_weight_quantizers.45 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.gate_up_proj_weight_quantizers.46 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.gate_up_proj_weight_quantizers.47 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.gate_up_proj_weight_quantizers.48 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.gate_up_proj_weight_quantizers.49 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.27e-01 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.gate_up_proj_weight_quantizers.50 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.gate_up_proj_weight_quantizers.51 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.gate_up_proj_weight_quantizers.52 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.gate_up_proj_weight_quantizers.53 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.gate_up_proj_weight_quantizers.54 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.02e-01 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.gate_up_proj_weight_quantizers.55 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.gate_up_proj_weight_quantizers.56 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.37e-01 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.gate_up_proj_weight_quantizers.57 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.gate_up_proj_weight_quantizers.58 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.gate_up_proj_weight_quantizers.59 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.gate_up_proj_weight_quantizers.60 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.gate_up_proj_weight_quantizers.61 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.gate_up_proj_weight_quantizers.62 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.gate_up_proj_weight_quantizers.63 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.44e-01 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.gate_up_proj_weight_quantizers.64 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.gate_up_proj_weight_quantizers.65 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.15e-01 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.gate_up_proj_weight_quantizers.66 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.gate_up_proj_weight_quantizers.67 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.gate_up_proj_weight_quantizers.68 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.88e-01 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.gate_up_proj_weight_quantizers.69 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.gate_up_proj_weight_quantizers.70 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.gate_up_proj_weight_quantizers.71 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.gate_up_proj_weight_quantizers.72 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.gate_up_proj_weight_quantizers.73 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.gate_up_proj_weight_quantizers.74 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.gate_up_proj_weight_quantizers.75 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=4.30e-01 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.gate_up_proj_weight_quantizers.76 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.gate_up_proj_weight_quantizers.77 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.gate_up_proj_weight_quantizers.78 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.gate_up_proj_weight_quantizers.79 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.53e-01 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.gate_up_proj_weight_quantizers.80 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.gate_up_proj_weight_quantizers.81 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.gate_up_proj_weight_quantizers.82 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.gate_up_proj_weight_quantizers.83 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.gate_up_proj_weight_quantizers.84 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.gate_up_proj_weight_quantizers.85 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.gate_up_proj_weight_quantizers.86 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.gate_up_proj_weight_quantizers.87 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.gate_up_proj_weight_quantizers.88 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.gate_up_proj_weight_quantizers.89 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.54e-01 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.gate_up_proj_weight_quantizers.90 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.gate_up_proj_weight_quantizers.91 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.26e-01 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.gate_up_proj_weight_quantizers.92 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.gate_up_proj_weight_quantizers.93 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.gate_up_proj_weight_quantizers.94 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.gate_up_proj_weight_quantizers.95 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.gate_up_proj_weight_quantizers.96 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.gate_up_proj_weight_quantizers.97 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.gate_up_proj_weight_quantizers.98 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.gate_up_proj_weight_quantizers.99 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.gate_up_proj_weight_quantizers.100 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.16e-01 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.gate_up_proj_weight_quantizers.101 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.29e-01 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.gate_up_proj_weight_quantizers.102 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.gate_up_proj_weight_quantizers.103 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.gate_up_proj_weight_quantizers.104 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.gate_up_proj_weight_quantizers.105 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.gate_up_proj_weight_quantizers.106 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.gate_up_proj_weight_quantizers.107 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.gate_up_proj_weight_quantizers.108 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.24e-01 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.gate_up_proj_weight_quantizers.109 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.gate_up_proj_weight_quantizers.110 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.gate_up_proj_weight_quantizers.111 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.gate_up_proj_weight_quantizers.112 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.gate_up_proj_weight_quantizers.113 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.gate_up_proj_weight_quantizers.114 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.25e-01 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.gate_up_proj_weight_quantizers.115 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.gate_up_proj_weight_quantizers.116 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.gate_up_proj_weight_quantizers.117 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.gate_up_proj_weight_quantizers.118 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.gate_up_proj_weight_quantizers.119 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.16e-01 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.gate_up_proj_weight_quantizers.120 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.gate_up_proj_weight_quantizers.121 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.gate_up_proj_weight_quantizers.122 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.gate_up_proj_weight_quantizers.123 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.gate_up_proj_weight_quantizers.124 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.gate_up_proj_weight_quantizers.125 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.24e-01 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.gate_up_proj_weight_quantizers.126 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.gate_up_proj_weight_quantizers.127 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.gate_up_proj_weight_quantizers.128 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.24e-01 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.gate_up_proj_weight_quantizers.129 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.32e-01 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.gate_up_proj_weight_quantizers.130 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.gate_up_proj_weight_quantizers.131 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.gate_up_proj_weight_quantizers.132 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.gate_up_proj_weight_quantizers.133 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.gate_up_proj_weight_quantizers.134 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.gate_up_proj_weight_quantizers.135 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.gate_up_proj_weight_quantizers.136 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.gate_up_proj_weight_quantizers.137 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.86e-01 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.gate_up_proj_weight_quantizers.138 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.gate_up_proj_weight_quantizers.139 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.gate_up_proj_weight_quantizers.140 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.gate_up_proj_weight_quantizers.141 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.gate_up_proj_weight_quantizers.142 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.gate_up_proj_weight_quantizers.143 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.gate_up_proj_weight_quantizers.144 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.gate_up_proj_weight_quantizers.145 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.gate_up_proj_weight_quantizers.146 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.gate_up_proj_weight_quantizers.147 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.gate_up_proj_weight_quantizers.148 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.gate_up_proj_weight_quantizers.149 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.gate_up_proj_weight_quantizers.150 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.gate_up_proj_weight_quantizers.151 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.gate_up_proj_weight_quantizers.152 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.gate_up_proj_weight_quantizers.153 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.gate_up_proj_weight_quantizers.154 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.gate_up_proj_weight_quantizers.155 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.56e-01 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.gate_up_proj_weight_quantizers.156 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.gate_up_proj_weight_quantizers.157 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.gate_up_proj_weight_quantizers.158 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.gate_up_proj_weight_quantizers.159 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.gate_up_proj_weight_quantizers.160 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.gate_up_proj_weight_quantizers.161 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.gate_up_proj_weight_quantizers.162 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.gate_up_proj_weight_quantizers.163 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.gate_up_proj_weight_quantizers.164 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.gate_up_proj_weight_quantizers.165 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.33e-01 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.gate_up_proj_weight_quantizers.166 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.gate_up_proj_weight_quantizers.167 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.gate_up_proj_weight_quantizers.168 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.gate_up_proj_weight_quantizers.169 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.22e-01 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.gate_up_proj_weight_quantizers.170 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.gate_up_proj_weight_quantizers.171 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.gate_up_proj_weight_quantizers.172 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.gate_up_proj_weight_quantizers.173 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.gate_up_proj_weight_quantizers.174 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.gate_up_proj_weight_quantizers.175 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.gate_up_proj_weight_quantizers.176 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.gate_up_proj_weight_quantizers.177 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.gate_up_proj_weight_quantizers.178 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.gate_up_proj_weight_quantizers.179 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.gate_up_proj_weight_quantizers.180 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.gate_up_proj_weight_quantizers.181 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.gate_up_proj_weight_quantizers.182 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.gate_up_proj_weight_quantizers.183 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.gate_up_proj_weight_quantizers.184 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.gate_up_proj_weight_quantizers.185 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.gate_up_proj_weight_quantizers.186 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.gate_up_proj_weight_quantizers.187 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.gate_up_proj_weight_quantizers.188 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.gate_up_proj_weight_quantizers.189 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.gate_up_proj_weight_quantizers.190 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.gate_up_proj_weight_quantizers.191 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.gate_up_proj_weight_quantizers.192 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.34e-01 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.gate_up_proj_weight_quantizers.193 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.gate_up_proj_weight_quantizers.194 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.gate_up_proj_weight_quantizers.195 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.gate_up_proj_weight_quantizers.196 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.gate_up_proj_weight_quantizers.197 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.gate_up_proj_weight_quantizers.198 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.gate_up_proj_weight_quantizers.199 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.10e-01 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.gate_up_proj_weight_quantizers.200 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.gate_up_proj_weight_quantizers.201 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.gate_up_proj_weight_quantizers.202 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.gate_up_proj_weight_quantizers.203 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.gate_up_proj_weight_quantizers.204 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.gate_up_proj_weight_quantizers.205 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.gate_up_proj_weight_quantizers.206 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.gate_up_proj_weight_quantizers.207 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.14e-01 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.gate_up_proj_weight_quantizers.208 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.gate_up_proj_weight_quantizers.209 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.gate_up_proj_weight_quantizers.210 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.gate_up_proj_weight_quantizers.211 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.64e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.gate_up_proj_weight_quantizers.212 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.gate_up_proj_weight_quantizers.213 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.gate_up_proj_weight_quantizers.214 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.gate_up_proj_weight_quantizers.215 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.gate_up_proj_weight_quantizers.216 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.74e-01 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.gate_up_proj_weight_quantizers.217 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.gate_up_proj_weight_quantizers.218 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.gate_up_proj_weight_quantizers.219 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.gate_up_proj_weight_quantizers.220 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.68e-01 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.gate_up_proj_weight_quantizers.221 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.88e-01 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.gate_up_proj_weight_quantizers.222 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.90e-01 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.gate_up_proj_weight_quantizers.223 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.gate_up_proj_weight_quantizers.224 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.gate_up_proj_weight_quantizers.225 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.gate_up_proj_weight_quantizers.226 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.gate_up_proj_weight_quantizers.227 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.gate_up_proj_weight_quantizers.228 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.gate_up_proj_weight_quantizers.229 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.gate_up_proj_weight_quantizers.230 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.gate_up_proj_weight_quantizers.231 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.gate_up_proj_weight_quantizers.232 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.gate_up_proj_weight_quantizers.233 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.gate_up_proj_weight_quantizers.234 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.gate_up_proj_weight_quantizers.235 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.38e-01 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.gate_up_proj_weight_quantizers.236 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.gate_up_proj_weight_quantizers.237 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.gate_up_proj_weight_quantizers.238 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.gate_up_proj_weight_quantizers.239 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.gate_up_proj_weight_quantizers.240 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.gate_up_proj_weight_quantizers.241 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.gate_up_proj_weight_quantizers.242 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.54e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.gate_up_proj_weight_quantizers.243 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.gate_up_proj_weight_quantizers.244 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.gate_up_proj_weight_quantizers.245 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.gate_up_proj_weight_quantizers.246 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.gate_up_proj_weight_quantizers.247 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.gate_up_proj_weight_quantizers.248 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.gate_up_proj_weight_quantizers.249 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.43e-01 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.gate_up_proj_weight_quantizers.250 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.gate_up_proj_weight_quantizers.251 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.gate_up_proj_weight_quantizers.252 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.gate_up_proj_weight_quantizers.253 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.gate_up_proj_weight_quantizers.254 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.gate_up_proj_weight_quantizers.255 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.down_proj_input_quantizer TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=6.12e+00 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.down_proj_weight_quantizers.0 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.down_proj_weight_quantizers.1 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.down_proj_weight_quantizers.2 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.down_proj_weight_quantizers.3 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.42e-01 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.down_proj_weight_quantizers.4 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.down_proj_weight_quantizers.5 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.down_proj_weight_quantizers.6 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.down_proj_weight_quantizers.7 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.down_proj_weight_quantizers.8 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.down_proj_weight_quantizers.9 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.64e-01 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.down_proj_weight_quantizers.10 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.down_proj_weight_quantizers.11 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.16e-01 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.down_proj_weight_quantizers.12 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.down_proj_weight_quantizers.13 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.down_proj_weight_quantizers.14 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.22e-01 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.down_proj_weight_quantizers.15 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.down_proj_weight_quantizers.16 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.down_proj_weight_quantizers.17 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.down_proj_weight_quantizers.18 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.down_proj_weight_quantizers.19 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.down_proj_weight_quantizers.20 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.down_proj_weight_quantizers.21 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.down_proj_weight_quantizers.22 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.down_proj_weight_quantizers.23 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.39e-01 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.down_proj_weight_quantizers.24 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.down_proj_weight_quantizers.25 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.down_proj_weight_quantizers.26 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.77e-01 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.down_proj_weight_quantizers.27 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.33e-01 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.down_proj_weight_quantizers.28 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.down_proj_weight_quantizers.29 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.56e-01 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.down_proj_weight_quantizers.30 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.down_proj_weight_quantizers.31 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.36e-01 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.down_proj_weight_quantizers.32 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.down_proj_weight_quantizers.33 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.down_proj_weight_quantizers.34 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.down_proj_weight_quantizers.35 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.down_proj_weight_quantizers.36 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.down_proj_weight_quantizers.37 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.down_proj_weight_quantizers.38 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.down_proj_weight_quantizers.39 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.down_proj_weight_quantizers.40 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.down_proj_weight_quantizers.41 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.down_proj_weight_quantizers.42 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.down_proj_weight_quantizers.43 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.down_proj_weight_quantizers.44 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.47e-01 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.down_proj_weight_quantizers.45 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.59e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.down_proj_weight_quantizers.46 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.down_proj_weight_quantizers.47 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.40e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.down_proj_weight_quantizers.48 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.down_proj_weight_quantizers.49 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.down_proj_weight_quantizers.50 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.90e-01 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.down_proj_weight_quantizers.51 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.down_proj_weight_quantizers.52 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.down_proj_weight_quantizers.53 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.16e-01 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.down_proj_weight_quantizers.54 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.down_proj_weight_quantizers.55 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.down_proj_weight_quantizers.56 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.79e-01 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.down_proj_weight_quantizers.57 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.down_proj_weight_quantizers.58 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.down_proj_weight_quantizers.59 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.down_proj_weight_quantizers.60 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.down_proj_weight_quantizers.61 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.16e-01 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.down_proj_weight_quantizers.62 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.down_proj_weight_quantizers.63 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.65e-01 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.down_proj_weight_quantizers.64 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.down_proj_weight_quantizers.65 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.down_proj_weight_quantizers.66 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.down_proj_weight_quantizers.67 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.down_proj_weight_quantizers.68 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.26e-01 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.down_proj_weight_quantizers.69 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.down_proj_weight_quantizers.70 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.down_proj_weight_quantizers.71 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.down_proj_weight_quantizers.72 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.down_proj_weight_quantizers.73 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.50e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.down_proj_weight_quantizers.74 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.down_proj_weight_quantizers.75 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.down_proj_weight_quantizers.76 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.down_proj_weight_quantizers.77 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.down_proj_weight_quantizers.78 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.down_proj_weight_quantizers.79 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.down_proj_weight_quantizers.80 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.down_proj_weight_quantizers.81 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.down_proj_weight_quantizers.82 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.down_proj_weight_quantizers.83 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.down_proj_weight_quantizers.84 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.down_proj_weight_quantizers.85 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.down_proj_weight_quantizers.86 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.77e-01 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.down_proj_weight_quantizers.87 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.40e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.down_proj_weight_quantizers.88 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.down_proj_weight_quantizers.89 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.56e-01 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.down_proj_weight_quantizers.90 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.down_proj_weight_quantizers.91 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.down_proj_weight_quantizers.92 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.64e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.down_proj_weight_quantizers.93 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.77e-01 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.down_proj_weight_quantizers.94 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.down_proj_weight_quantizers.95 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.80e-01 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.down_proj_weight_quantizers.96 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.down_proj_weight_quantizers.97 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.down_proj_weight_quantizers.98 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.down_proj_weight_quantizers.99 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.down_proj_weight_quantizers.100 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.79e-01 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.down_proj_weight_quantizers.101 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.38e-01 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.down_proj_weight_quantizers.102 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.down_proj_weight_quantizers.103 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.down_proj_weight_quantizers.104 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.down_proj_weight_quantizers.105 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.down_proj_weight_quantizers.106 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.54e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.down_proj_weight_quantizers.107 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.down_proj_weight_quantizers.108 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.19e-01 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.down_proj_weight_quantizers.109 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.down_proj_weight_quantizers.110 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.down_proj_weight_quantizers.111 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.down_proj_weight_quantizers.112 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.down_proj_weight_quantizers.113 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.down_proj_weight_quantizers.114 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.38e-01 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.down_proj_weight_quantizers.115 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.down_proj_weight_quantizers.116 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.down_proj_weight_quantizers.117 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.down_proj_weight_quantizers.118 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.down_proj_weight_quantizers.119 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.46e-01 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.down_proj_weight_quantizers.120 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.down_proj_weight_quantizers.121 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.down_proj_weight_quantizers.122 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.down_proj_weight_quantizers.123 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.down_proj_weight_quantizers.124 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.down_proj_weight_quantizers.125 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.75e-01 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.down_proj_weight_quantizers.126 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.down_proj_weight_quantizers.127 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.down_proj_weight_quantizers.128 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.down_proj_weight_quantizers.129 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.down_proj_weight_quantizers.130 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.down_proj_weight_quantizers.131 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.down_proj_weight_quantizers.132 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.26e-01 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.down_proj_weight_quantizers.133 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.down_proj_weight_quantizers.134 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.down_proj_weight_quantizers.135 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.down_proj_weight_quantizers.136 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.down_proj_weight_quantizers.137 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.41e-01 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.down_proj_weight_quantizers.138 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.down_proj_weight_quantizers.139 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.down_proj_weight_quantizers.140 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.down_proj_weight_quantizers.141 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.down_proj_weight_quantizers.142 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.down_proj_weight_quantizers.143 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.43e-01 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.down_proj_weight_quantizers.144 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.down_proj_weight_quantizers.145 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.down_proj_weight_quantizers.146 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.down_proj_weight_quantizers.147 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.down_proj_weight_quantizers.148 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.down_proj_weight_quantizers.149 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.down_proj_weight_quantizers.150 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.down_proj_weight_quantizers.151 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.down_proj_weight_quantizers.152 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.50e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.down_proj_weight_quantizers.153 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.down_proj_weight_quantizers.154 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.54e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.down_proj_weight_quantizers.155 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.down_proj_weight_quantizers.156 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.down_proj_weight_quantizers.157 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.down_proj_weight_quantizers.158 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.down_proj_weight_quantizers.159 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.down_proj_weight_quantizers.160 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.18e-01 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.down_proj_weight_quantizers.161 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.down_proj_weight_quantizers.162 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.27e-01 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.down_proj_weight_quantizers.163 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.down_proj_weight_quantizers.164 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.down_proj_weight_quantizers.165 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.83e-01 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.down_proj_weight_quantizers.166 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.down_proj_weight_quantizers.167 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.64e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.down_proj_weight_quantizers.168 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.down_proj_weight_quantizers.169 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.down_proj_weight_quantizers.170 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.down_proj_weight_quantizers.171 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.down_proj_weight_quantizers.172 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.down_proj_weight_quantizers.173 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.down_proj_weight_quantizers.174 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.down_proj_weight_quantizers.175 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.down_proj_weight_quantizers.176 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.20e-01 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.down_proj_weight_quantizers.177 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.down_proj_weight_quantizers.178 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.down_proj_weight_quantizers.179 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.down_proj_weight_quantizers.180 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.down_proj_weight_quantizers.181 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.59e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.down_proj_weight_quantizers.182 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.down_proj_weight_quantizers.183 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.down_proj_weight_quantizers.184 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.88e-01 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.down_proj_weight_quantizers.185 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.down_proj_weight_quantizers.186 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.down_proj_weight_quantizers.187 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.down_proj_weight_quantizers.188 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.59e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.down_proj_weight_quantizers.189 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.down_proj_weight_quantizers.190 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.down_proj_weight_quantizers.191 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.down_proj_weight_quantizers.192 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.19e-01 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.down_proj_weight_quantizers.193 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.down_proj_weight_quantizers.194 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.down_proj_weight_quantizers.195 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.down_proj_weight_quantizers.196 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.down_proj_weight_quantizers.197 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.down_proj_weight_quantizers.198 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.59e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.down_proj_weight_quantizers.199 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.11e-01 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.down_proj_weight_quantizers.200 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.27e-01 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.down_proj_weight_quantizers.201 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.down_proj_weight_quantizers.202 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.down_proj_weight_quantizers.203 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.down_proj_weight_quantizers.204 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.down_proj_weight_quantizers.205 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.40e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.down_proj_weight_quantizers.206 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.down_proj_weight_quantizers.207 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.down_proj_weight_quantizers.208 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.down_proj_weight_quantizers.209 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.down_proj_weight_quantizers.210 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.down_proj_weight_quantizers.211 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.down_proj_weight_quantizers.212 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.down_proj_weight_quantizers.213 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.54e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.down_proj_weight_quantizers.214 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.down_proj_weight_quantizers.215 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.down_proj_weight_quantizers.216 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.23e-01 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.down_proj_weight_quantizers.217 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.down_proj_weight_quantizers.218 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.down_proj_weight_quantizers.219 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.down_proj_weight_quantizers.220 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.74e-01 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.down_proj_weight_quantizers.221 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.17e-01 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.down_proj_weight_quantizers.222 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.down_proj_weight_quantizers.223 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.down_proj_weight_quantizers.224 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.90e-01 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.down_proj_weight_quantizers.225 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.down_proj_weight_quantizers.226 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.down_proj_weight_quantizers.227 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.down_proj_weight_quantizers.228 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.down_proj_weight_quantizers.229 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.down_proj_weight_quantizers.230 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.54e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.down_proj_weight_quantizers.231 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.down_proj_weight_quantizers.232 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.down_proj_weight_quantizers.233 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.down_proj_weight_quantizers.234 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.64e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.down_proj_weight_quantizers.235 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.31e-01 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.down_proj_weight_quantizers.236 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.down_proj_weight_quantizers.237 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.down_proj_weight_quantizers.238 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.down_proj_weight_quantizers.239 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.down_proj_weight_quantizers.240 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.down_proj_weight_quantizers.241 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.64e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.down_proj_weight_quantizers.242 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.down_proj_weight_quantizers.243 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.down_proj_weight_quantizers.244 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.down_proj_weight_quantizers.245 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.down_proj_weight_quantizers.246 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.35e-01 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.down_proj_weight_quantizers.247 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.down_proj_weight_quantizers.248 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.down_proj_weight_quantizers.249 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.17e-01 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.down_proj_weight_quantizers.250 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.22e-01 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.down_proj_weight_quantizers.251 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.down_proj_weight_quantizers.252 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.down_proj_weight_quantizers.253 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.down_proj_weight_quantizers.254 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.14e-01 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.down_proj_weight_quantizers.255 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.shared_experts.gate_proj.input_quantizer TensorQuantizer(disabled) +model.layers.23.mlp.shared_experts.gate_proj.output_quantizer TensorQuantizer(disabled) +model.layers.23.mlp.shared_experts.gate_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.23.mlp.shared_experts.up_proj.input_quantizer TensorQuantizer(disabled) +model.layers.23.mlp.shared_experts.up_proj.output_quantizer TensorQuantizer(disabled) +model.layers.23.mlp.shared_experts.up_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.23.mlp.shared_experts.down_proj.input_quantizer TensorQuantizer(disabled) +model.layers.23.mlp.shared_experts.down_proj.output_quantizer TensorQuantizer(disabled) +model.layers.23.mlp.shared_experts.down_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.24.self_attn.q_a_proj.input_quantizer TensorQuantizer(disabled) +model.layers.24.self_attn.q_a_proj.output_quantizer TensorQuantizer(disabled) +model.layers.24.self_attn.q_a_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.24.self_attn.q_b_proj.input_quantizer TensorQuantizer(disabled) +model.layers.24.self_attn.q_b_proj.output_quantizer TensorQuantizer(disabled) +model.layers.24.self_attn.q_b_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.24.self_attn.kv_a_proj_with_mqa.input_quantizer TensorQuantizer(disabled) +model.layers.24.self_attn.kv_a_proj_with_mqa.output_quantizer TensorQuantizer(disabled) +model.layers.24.self_attn.kv_a_proj_with_mqa.weight_quantizer TensorQuantizer(disabled) +model.layers.24.self_attn.kv_b_proj.input_quantizer TensorQuantizer(disabled) +model.layers.24.self_attn.kv_b_proj.output_quantizer TensorQuantizer(disabled) +model.layers.24.self_attn.kv_b_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.24.self_attn.o_proj.input_quantizer TensorQuantizer(disabled) +model.layers.24.self_attn.o_proj.output_quantizer TensorQuantizer(disabled) +model.layers.24.self_attn.o_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.24.self_attn.q_bmm_quantizer TensorQuantizer(disabled) +model.layers.24.self_attn.k_bmm_quantizer TensorQuantizer((4, 3) bit fake per-tensor amax=4.48e+02(const) calibrator=MaxCalibrator quant) +model.layers.24.self_attn.v_bmm_quantizer TensorQuantizer((4, 3) bit fake per-tensor amax=4.48e+02(const) calibrator=MaxCalibrator quant) +model.layers.24.self_attn.p_bmm_quantizer TensorQuantizer(disabled) +model.layers.24.mlp.experts.gate_up_proj_input_quantizer TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.91e-01 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.gate_up_proj_weight_quantizers.0 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.gate_up_proj_weight_quantizers.1 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.gate_up_proj_weight_quantizers.2 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.16e-01 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.gate_up_proj_weight_quantizers.3 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.gate_up_proj_weight_quantizers.4 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.25e-01 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.gate_up_proj_weight_quantizers.5 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.gate_up_proj_weight_quantizers.6 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.gate_up_proj_weight_quantizers.7 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.gate_up_proj_weight_quantizers.8 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.gate_up_proj_weight_quantizers.9 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.gate_up_proj_weight_quantizers.10 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.gate_up_proj_weight_quantizers.11 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.gate_up_proj_weight_quantizers.12 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.gate_up_proj_weight_quantizers.13 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.gate_up_proj_weight_quantizers.14 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.gate_up_proj_weight_quantizers.15 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.gate_up_proj_weight_quantizers.16 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.20e-01 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.gate_up_proj_weight_quantizers.17 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.gate_up_proj_weight_quantizers.18 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.59e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.gate_up_proj_weight_quantizers.19 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.gate_up_proj_weight_quantizers.20 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.gate_up_proj_weight_quantizers.21 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.13e-01 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.gate_up_proj_weight_quantizers.22 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.gate_up_proj_weight_quantizers.23 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.gate_up_proj_weight_quantizers.24 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.gate_up_proj_weight_quantizers.25 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.gate_up_proj_weight_quantizers.26 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.gate_up_proj_weight_quantizers.27 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.gate_up_proj_weight_quantizers.28 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.gate_up_proj_weight_quantizers.29 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.gate_up_proj_weight_quantizers.30 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.28e-01 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.gate_up_proj_weight_quantizers.31 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.gate_up_proj_weight_quantizers.32 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.gate_up_proj_weight_quantizers.33 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.72e-01 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.gate_up_proj_weight_quantizers.34 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.gate_up_proj_weight_quantizers.35 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.gate_up_proj_weight_quantizers.36 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.gate_up_proj_weight_quantizers.37 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.gate_up_proj_weight_quantizers.38 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.gate_up_proj_weight_quantizers.39 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.64e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.gate_up_proj_weight_quantizers.40 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.gate_up_proj_weight_quantizers.41 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.gate_up_proj_weight_quantizers.42 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.gate_up_proj_weight_quantizers.43 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.gate_up_proj_weight_quantizers.44 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.56e-01 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.gate_up_proj_weight_quantizers.45 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.gate_up_proj_weight_quantizers.46 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.gate_up_proj_weight_quantizers.47 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.97e-01 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.gate_up_proj_weight_quantizers.48 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.50e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.gate_up_proj_weight_quantizers.49 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.gate_up_proj_weight_quantizers.50 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.gate_up_proj_weight_quantizers.51 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.gate_up_proj_weight_quantizers.52 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.gate_up_proj_weight_quantizers.53 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.57e-01 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.gate_up_proj_weight_quantizers.54 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.gate_up_proj_weight_quantizers.55 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.gate_up_proj_weight_quantizers.56 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.gate_up_proj_weight_quantizers.57 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.gate_up_proj_weight_quantizers.58 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.gate_up_proj_weight_quantizers.59 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.gate_up_proj_weight_quantizers.60 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.gate_up_proj_weight_quantizers.61 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.gate_up_proj_weight_quantizers.62 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.gate_up_proj_weight_quantizers.63 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.gate_up_proj_weight_quantizers.64 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.gate_up_proj_weight_quantizers.65 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.gate_up_proj_weight_quantizers.66 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.gate_up_proj_weight_quantizers.67 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.gate_up_proj_weight_quantizers.68 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.gate_up_proj_weight_quantizers.69 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.gate_up_proj_weight_quantizers.70 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.gate_up_proj_weight_quantizers.71 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.44e-01 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.gate_up_proj_weight_quantizers.72 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.gate_up_proj_weight_quantizers.73 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.gate_up_proj_weight_quantizers.74 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.gate_up_proj_weight_quantizers.75 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.gate_up_proj_weight_quantizers.76 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.gate_up_proj_weight_quantizers.77 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.gate_up_proj_weight_quantizers.78 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.gate_up_proj_weight_quantizers.79 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.gate_up_proj_weight_quantizers.80 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.gate_up_proj_weight_quantizers.81 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.gate_up_proj_weight_quantizers.82 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.gate_up_proj_weight_quantizers.83 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.gate_up_proj_weight_quantizers.84 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.30e-01 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.gate_up_proj_weight_quantizers.85 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.gate_up_proj_weight_quantizers.86 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.gate_up_proj_weight_quantizers.87 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.gate_up_proj_weight_quantizers.88 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.gate_up_proj_weight_quantizers.89 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.gate_up_proj_weight_quantizers.90 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.05e-01 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.gate_up_proj_weight_quantizers.91 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.28e-01 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.gate_up_proj_weight_quantizers.92 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.gate_up_proj_weight_quantizers.93 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.gate_up_proj_weight_quantizers.94 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.gate_up_proj_weight_quantizers.95 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.32e-01 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.gate_up_proj_weight_quantizers.96 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.gate_up_proj_weight_quantizers.97 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.gate_up_proj_weight_quantizers.98 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.gate_up_proj_weight_quantizers.99 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.29e-01 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.gate_up_proj_weight_quantizers.100 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.gate_up_proj_weight_quantizers.101 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.gate_up_proj_weight_quantizers.102 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.gate_up_proj_weight_quantizers.103 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.gate_up_proj_weight_quantizers.104 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.gate_up_proj_weight_quantizers.105 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.gate_up_proj_weight_quantizers.106 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.gate_up_proj_weight_quantizers.107 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.gate_up_proj_weight_quantizers.108 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.gate_up_proj_weight_quantizers.109 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.gate_up_proj_weight_quantizers.110 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.gate_up_proj_weight_quantizers.111 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.gate_up_proj_weight_quantizers.112 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.gate_up_proj_weight_quantizers.113 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.gate_up_proj_weight_quantizers.114 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.gate_up_proj_weight_quantizers.115 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.gate_up_proj_weight_quantizers.116 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.15e-01 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.gate_up_proj_weight_quantizers.117 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.gate_up_proj_weight_quantizers.118 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.gate_up_proj_weight_quantizers.119 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.47e-01 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.gate_up_proj_weight_quantizers.120 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.gate_up_proj_weight_quantizers.121 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.gate_up_proj_weight_quantizers.122 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.gate_up_proj_weight_quantizers.123 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.21e-01 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.gate_up_proj_weight_quantizers.124 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.gate_up_proj_weight_quantizers.125 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.gate_up_proj_weight_quantizers.126 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.gate_up_proj_weight_quantizers.127 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.gate_up_proj_weight_quantizers.128 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.gate_up_proj_weight_quantizers.129 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.gate_up_proj_weight_quantizers.130 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.gate_up_proj_weight_quantizers.131 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.gate_up_proj_weight_quantizers.132 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.47e-01 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.gate_up_proj_weight_quantizers.133 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.gate_up_proj_weight_quantizers.134 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.gate_up_proj_weight_quantizers.135 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.34e-01 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.gate_up_proj_weight_quantizers.136 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.20e-01 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.gate_up_proj_weight_quantizers.137 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.gate_up_proj_weight_quantizers.138 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.22e-01 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.gate_up_proj_weight_quantizers.139 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.gate_up_proj_weight_quantizers.140 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.gate_up_proj_weight_quantizers.141 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.gate_up_proj_weight_quantizers.142 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.gate_up_proj_weight_quantizers.143 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.gate_up_proj_weight_quantizers.144 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.gate_up_proj_weight_quantizers.145 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.gate_up_proj_weight_quantizers.146 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.gate_up_proj_weight_quantizers.147 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.gate_up_proj_weight_quantizers.148 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.gate_up_proj_weight_quantizers.149 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.gate_up_proj_weight_quantizers.150 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.52e-01 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.gate_up_proj_weight_quantizers.151 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.gate_up_proj_weight_quantizers.152 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.gate_up_proj_weight_quantizers.153 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.gate_up_proj_weight_quantizers.154 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.gate_up_proj_weight_quantizers.155 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.gate_up_proj_weight_quantizers.156 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.gate_up_proj_weight_quantizers.157 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.gate_up_proj_weight_quantizers.158 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.gate_up_proj_weight_quantizers.159 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.gate_up_proj_weight_quantizers.160 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.gate_up_proj_weight_quantizers.161 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.gate_up_proj_weight_quantizers.162 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.gate_up_proj_weight_quantizers.163 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.gate_up_proj_weight_quantizers.164 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.gate_up_proj_weight_quantizers.165 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.gate_up_proj_weight_quantizers.166 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.gate_up_proj_weight_quantizers.167 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.gate_up_proj_weight_quantizers.168 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.gate_up_proj_weight_quantizers.169 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.23e-01 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.gate_up_proj_weight_quantizers.170 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.gate_up_proj_weight_quantizers.171 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.gate_up_proj_weight_quantizers.172 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.gate_up_proj_weight_quantizers.173 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.gate_up_proj_weight_quantizers.174 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.gate_up_proj_weight_quantizers.175 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.gate_up_proj_weight_quantizers.176 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.gate_up_proj_weight_quantizers.177 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.gate_up_proj_weight_quantizers.178 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.gate_up_proj_weight_quantizers.179 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.gate_up_proj_weight_quantizers.180 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.gate_up_proj_weight_quantizers.181 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.59e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.gate_up_proj_weight_quantizers.182 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.gate_up_proj_weight_quantizers.183 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.17e-01 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.gate_up_proj_weight_quantizers.184 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.gate_up_proj_weight_quantizers.185 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.gate_up_proj_weight_quantizers.186 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.30e-01 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.gate_up_proj_weight_quantizers.187 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.gate_up_proj_weight_quantizers.188 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.44e-01 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.gate_up_proj_weight_quantizers.189 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.gate_up_proj_weight_quantizers.190 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.gate_up_proj_weight_quantizers.191 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.gate_up_proj_weight_quantizers.192 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.gate_up_proj_weight_quantizers.193 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.gate_up_proj_weight_quantizers.194 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.gate_up_proj_weight_quantizers.195 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.gate_up_proj_weight_quantizers.196 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.gate_up_proj_weight_quantizers.197 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.gate_up_proj_weight_quantizers.198 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.gate_up_proj_weight_quantizers.199 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.gate_up_proj_weight_quantizers.200 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.gate_up_proj_weight_quantizers.201 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.gate_up_proj_weight_quantizers.202 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.gate_up_proj_weight_quantizers.203 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.gate_up_proj_weight_quantizers.204 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.gate_up_proj_weight_quantizers.205 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.gate_up_proj_weight_quantizers.206 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.gate_up_proj_weight_quantizers.207 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.gate_up_proj_weight_quantizers.208 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.gate_up_proj_weight_quantizers.209 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.gate_up_proj_weight_quantizers.210 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.gate_up_proj_weight_quantizers.211 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.gate_up_proj_weight_quantizers.212 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.gate_up_proj_weight_quantizers.213 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.gate_up_proj_weight_quantizers.214 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.gate_up_proj_weight_quantizers.215 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.gate_up_proj_weight_quantizers.216 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.gate_up_proj_weight_quantizers.217 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.gate_up_proj_weight_quantizers.218 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.gate_up_proj_weight_quantizers.219 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.gate_up_proj_weight_quantizers.220 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.gate_up_proj_weight_quantizers.221 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.gate_up_proj_weight_quantizers.222 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.gate_up_proj_weight_quantizers.223 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.gate_up_proj_weight_quantizers.224 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.gate_up_proj_weight_quantizers.225 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.gate_up_proj_weight_quantizers.226 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.gate_up_proj_weight_quantizers.227 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.19e-01 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.gate_up_proj_weight_quantizers.228 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.gate_up_proj_weight_quantizers.229 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.gate_up_proj_weight_quantizers.230 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.64e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.gate_up_proj_weight_quantizers.231 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.gate_up_proj_weight_quantizers.232 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.gate_up_proj_weight_quantizers.233 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.gate_up_proj_weight_quantizers.234 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.gate_up_proj_weight_quantizers.235 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.gate_up_proj_weight_quantizers.236 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.gate_up_proj_weight_quantizers.237 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.48e-01 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.gate_up_proj_weight_quantizers.238 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.gate_up_proj_weight_quantizers.239 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.gate_up_proj_weight_quantizers.240 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.gate_up_proj_weight_quantizers.241 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.gate_up_proj_weight_quantizers.242 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.gate_up_proj_weight_quantizers.243 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.gate_up_proj_weight_quantizers.244 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.gate_up_proj_weight_quantizers.245 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.gate_up_proj_weight_quantizers.246 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.gate_up_proj_weight_quantizers.247 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.gate_up_proj_weight_quantizers.248 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.15e-01 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.gate_up_proj_weight_quantizers.249 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.gate_up_proj_weight_quantizers.250 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.66e-01 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.gate_up_proj_weight_quantizers.251 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.gate_up_proj_weight_quantizers.252 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.gate_up_proj_weight_quantizers.253 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.gate_up_proj_weight_quantizers.254 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.gate_up_proj_weight_quantizers.255 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.down_proj_input_quantizer TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=6.09e+00 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.down_proj_weight_quantizers.0 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.down_proj_weight_quantizers.1 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.down_proj_weight_quantizers.2 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.down_proj_weight_quantizers.3 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.down_proj_weight_quantizers.4 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.down_proj_weight_quantizers.5 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.down_proj_weight_quantizers.6 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.54e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.down_proj_weight_quantizers.7 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.50e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.down_proj_weight_quantizers.8 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.down_proj_weight_quantizers.9 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.down_proj_weight_quantizers.10 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.down_proj_weight_quantizers.11 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.down_proj_weight_quantizers.12 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.down_proj_weight_quantizers.13 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.down_proj_weight_quantizers.14 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.down_proj_weight_quantizers.15 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.down_proj_weight_quantizers.16 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.98e-01 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.down_proj_weight_quantizers.17 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.50e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.down_proj_weight_quantizers.18 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.down_proj_weight_quantizers.19 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.down_proj_weight_quantizers.20 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.down_proj_weight_quantizers.21 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.45e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.down_proj_weight_quantizers.22 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.down_proj_weight_quantizers.23 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.40e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.down_proj_weight_quantizers.24 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.down_proj_weight_quantizers.25 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.down_proj_weight_quantizers.26 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.down_proj_weight_quantizers.27 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.down_proj_weight_quantizers.28 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.down_proj_weight_quantizers.29 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.down_proj_weight_quantizers.30 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.down_proj_weight_quantizers.31 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.down_proj_weight_quantizers.32 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.64e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.down_proj_weight_quantizers.33 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.down_proj_weight_quantizers.34 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.down_proj_weight_quantizers.35 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.54e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.down_proj_weight_quantizers.36 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.down_proj_weight_quantizers.37 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.down_proj_weight_quantizers.38 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.down_proj_weight_quantizers.39 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.down_proj_weight_quantizers.40 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.59e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.down_proj_weight_quantizers.41 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.30e-01 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.down_proj_weight_quantizers.42 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.down_proj_weight_quantizers.43 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.23e-01 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.down_proj_weight_quantizers.44 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.65e-01 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.down_proj_weight_quantizers.45 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.down_proj_weight_quantizers.46 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.down_proj_weight_quantizers.47 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.92e-01 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.down_proj_weight_quantizers.48 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.down_proj_weight_quantizers.49 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.down_proj_weight_quantizers.50 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.down_proj_weight_quantizers.51 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.45e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.down_proj_weight_quantizers.52 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.down_proj_weight_quantizers.53 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.57e-01 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.down_proj_weight_quantizers.54 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.28e-01 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.down_proj_weight_quantizers.55 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.down_proj_weight_quantizers.56 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.down_proj_weight_quantizers.57 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.down_proj_weight_quantizers.58 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.down_proj_weight_quantizers.59 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.down_proj_weight_quantizers.60 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.down_proj_weight_quantizers.61 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.down_proj_weight_quantizers.62 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.64e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.down_proj_weight_quantizers.63 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.45e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.down_proj_weight_quantizers.64 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.down_proj_weight_quantizers.65 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.down_proj_weight_quantizers.66 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.down_proj_weight_quantizers.67 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.down_proj_weight_quantizers.68 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.down_proj_weight_quantizers.69 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.down_proj_weight_quantizers.70 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.down_proj_weight_quantizers.71 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.21e-01 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.down_proj_weight_quantizers.72 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.down_proj_weight_quantizers.73 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.34e-01 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.down_proj_weight_quantizers.74 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.down_proj_weight_quantizers.75 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.down_proj_weight_quantizers.76 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.down_proj_weight_quantizers.77 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.54e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.down_proj_weight_quantizers.78 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.down_proj_weight_quantizers.79 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.down_proj_weight_quantizers.80 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.down_proj_weight_quantizers.81 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.down_proj_weight_quantizers.82 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.down_proj_weight_quantizers.83 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.23e-01 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.down_proj_weight_quantizers.84 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.22e-01 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.down_proj_weight_quantizers.85 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.down_proj_weight_quantizers.86 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.down_proj_weight_quantizers.87 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.down_proj_weight_quantizers.88 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.down_proj_weight_quantizers.89 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.down_proj_weight_quantizers.90 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.38e-01 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.down_proj_weight_quantizers.91 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.down_proj_weight_quantizers.92 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.down_proj_weight_quantizers.93 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.down_proj_weight_quantizers.94 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.down_proj_weight_quantizers.95 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.down_proj_weight_quantizers.96 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.down_proj_weight_quantizers.97 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.down_proj_weight_quantizers.98 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.down_proj_weight_quantizers.99 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.down_proj_weight_quantizers.100 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.down_proj_weight_quantizers.101 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.down_proj_weight_quantizers.102 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.down_proj_weight_quantizers.103 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.down_proj_weight_quantizers.104 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.down_proj_weight_quantizers.105 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.down_proj_weight_quantizers.106 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.down_proj_weight_quantizers.107 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.down_proj_weight_quantizers.108 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.down_proj_weight_quantizers.109 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.down_proj_weight_quantizers.110 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.down_proj_weight_quantizers.111 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.down_proj_weight_quantizers.112 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.down_proj_weight_quantizers.113 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.70e-01 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.down_proj_weight_quantizers.114 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.down_proj_weight_quantizers.115 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.down_proj_weight_quantizers.116 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.down_proj_weight_quantizers.117 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.down_proj_weight_quantizers.118 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.28e-01 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.down_proj_weight_quantizers.119 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.down_proj_weight_quantizers.120 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.down_proj_weight_quantizers.121 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.50e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.down_proj_weight_quantizers.122 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.down_proj_weight_quantizers.123 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.59e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.down_proj_weight_quantizers.124 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.down_proj_weight_quantizers.125 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.59e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.down_proj_weight_quantizers.126 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.down_proj_weight_quantizers.127 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.down_proj_weight_quantizers.128 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.down_proj_weight_quantizers.129 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.64e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.down_proj_weight_quantizers.130 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.down_proj_weight_quantizers.131 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.down_proj_weight_quantizers.132 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.76e-01 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.down_proj_weight_quantizers.133 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.down_proj_weight_quantizers.134 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.down_proj_weight_quantizers.135 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.87e-01 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.down_proj_weight_quantizers.136 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.25e-01 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.down_proj_weight_quantizers.137 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.down_proj_weight_quantizers.138 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.down_proj_weight_quantizers.139 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.down_proj_weight_quantizers.140 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.down_proj_weight_quantizers.141 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.down_proj_weight_quantizers.142 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.down_proj_weight_quantizers.143 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.down_proj_weight_quantizers.144 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.down_proj_weight_quantizers.145 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.down_proj_weight_quantizers.146 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.down_proj_weight_quantizers.147 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.down_proj_weight_quantizers.148 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.down_proj_weight_quantizers.149 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.down_proj_weight_quantizers.150 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.35e-01 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.down_proj_weight_quantizers.151 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.68e-01 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.down_proj_weight_quantizers.152 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.down_proj_weight_quantizers.153 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.15e-01 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.down_proj_weight_quantizers.154 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.down_proj_weight_quantizers.155 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.down_proj_weight_quantizers.156 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.64e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.down_proj_weight_quantizers.157 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.down_proj_weight_quantizers.158 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.down_proj_weight_quantizers.159 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.down_proj_weight_quantizers.160 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.down_proj_weight_quantizers.161 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.down_proj_weight_quantizers.162 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.45e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.down_proj_weight_quantizers.163 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.down_proj_weight_quantizers.164 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.down_proj_weight_quantizers.165 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.down_proj_weight_quantizers.166 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.28e-01 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.down_proj_weight_quantizers.167 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.16e-01 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.down_proj_weight_quantizers.168 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.down_proj_weight_quantizers.169 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.59e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.down_proj_weight_quantizers.170 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.85e-01 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.down_proj_weight_quantizers.171 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.83e-01 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.down_proj_weight_quantizers.172 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.down_proj_weight_quantizers.173 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.down_proj_weight_quantizers.174 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.down_proj_weight_quantizers.175 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.64e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.down_proj_weight_quantizers.176 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.down_proj_weight_quantizers.177 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.down_proj_weight_quantizers.178 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.down_proj_weight_quantizers.179 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.down_proj_weight_quantizers.180 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.down_proj_weight_quantizers.181 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.down_proj_weight_quantizers.182 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.down_proj_weight_quantizers.183 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.down_proj_weight_quantizers.184 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.down_proj_weight_quantizers.185 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.down_proj_weight_quantizers.186 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.47e-01 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.down_proj_weight_quantizers.187 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.down_proj_weight_quantizers.188 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.down_proj_weight_quantizers.189 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.down_proj_weight_quantizers.190 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.down_proj_weight_quantizers.191 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.down_proj_weight_quantizers.192 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.down_proj_weight_quantizers.193 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.61e-01 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.down_proj_weight_quantizers.194 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.down_proj_weight_quantizers.195 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.down_proj_weight_quantizers.196 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.down_proj_weight_quantizers.197 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.down_proj_weight_quantizers.198 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.down_proj_weight_quantizers.199 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.down_proj_weight_quantizers.200 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.down_proj_weight_quantizers.201 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.down_proj_weight_quantizers.202 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.down_proj_weight_quantizers.203 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.down_proj_weight_quantizers.204 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.down_proj_weight_quantizers.205 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.down_proj_weight_quantizers.206 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.down_proj_weight_quantizers.207 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.down_proj_weight_quantizers.208 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.54e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.down_proj_weight_quantizers.209 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.59e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.down_proj_weight_quantizers.210 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.down_proj_weight_quantizers.211 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.down_proj_weight_quantizers.212 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.59e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.down_proj_weight_quantizers.213 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.down_proj_weight_quantizers.214 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.down_proj_weight_quantizers.215 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.down_proj_weight_quantizers.216 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.down_proj_weight_quantizers.217 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.35e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.down_proj_weight_quantizers.218 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.down_proj_weight_quantizers.219 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.down_proj_weight_quantizers.220 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.down_proj_weight_quantizers.221 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.down_proj_weight_quantizers.222 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.down_proj_weight_quantizers.223 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.down_proj_weight_quantizers.224 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.down_proj_weight_quantizers.225 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.down_proj_weight_quantizers.226 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.down_proj_weight_quantizers.227 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.down_proj_weight_quantizers.228 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.down_proj_weight_quantizers.229 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.16e-01 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.down_proj_weight_quantizers.230 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.down_proj_weight_quantizers.231 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.down_proj_weight_quantizers.232 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.down_proj_weight_quantizers.233 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.down_proj_weight_quantizers.234 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.down_proj_weight_quantizers.235 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.down_proj_weight_quantizers.236 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.down_proj_weight_quantizers.237 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.down_proj_weight_quantizers.238 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.59e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.down_proj_weight_quantizers.239 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.down_proj_weight_quantizers.240 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.down_proj_weight_quantizers.241 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.down_proj_weight_quantizers.242 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.down_proj_weight_quantizers.243 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.down_proj_weight_quantizers.244 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.down_proj_weight_quantizers.245 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.down_proj_weight_quantizers.246 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.down_proj_weight_quantizers.247 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.down_proj_weight_quantizers.248 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.31e-01 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.down_proj_weight_quantizers.249 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.down_proj_weight_quantizers.250 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.48e-01 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.down_proj_weight_quantizers.251 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.54e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.down_proj_weight_quantizers.252 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.down_proj_weight_quantizers.253 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.down_proj_weight_quantizers.254 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.down_proj_weight_quantizers.255 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.shared_experts.gate_proj.input_quantizer TensorQuantizer(disabled) +model.layers.24.mlp.shared_experts.gate_proj.output_quantizer TensorQuantizer(disabled) +model.layers.24.mlp.shared_experts.gate_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.24.mlp.shared_experts.up_proj.input_quantizer TensorQuantizer(disabled) +model.layers.24.mlp.shared_experts.up_proj.output_quantizer TensorQuantizer(disabled) +model.layers.24.mlp.shared_experts.up_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.24.mlp.shared_experts.down_proj.input_quantizer TensorQuantizer(disabled) +model.layers.24.mlp.shared_experts.down_proj.output_quantizer TensorQuantizer(disabled) +model.layers.24.mlp.shared_experts.down_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.25.self_attn.q_a_proj.input_quantizer TensorQuantizer(disabled) +model.layers.25.self_attn.q_a_proj.output_quantizer TensorQuantizer(disabled) +model.layers.25.self_attn.q_a_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.25.self_attn.q_b_proj.input_quantizer TensorQuantizer(disabled) +model.layers.25.self_attn.q_b_proj.output_quantizer TensorQuantizer(disabled) +model.layers.25.self_attn.q_b_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.25.self_attn.kv_a_proj_with_mqa.input_quantizer TensorQuantizer(disabled) +model.layers.25.self_attn.kv_a_proj_with_mqa.output_quantizer TensorQuantizer(disabled) +model.layers.25.self_attn.kv_a_proj_with_mqa.weight_quantizer TensorQuantizer(disabled) +model.layers.25.self_attn.kv_b_proj.input_quantizer TensorQuantizer(disabled) +model.layers.25.self_attn.kv_b_proj.output_quantizer TensorQuantizer(disabled) +model.layers.25.self_attn.kv_b_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.25.self_attn.o_proj.input_quantizer TensorQuantizer(disabled) +model.layers.25.self_attn.o_proj.output_quantizer TensorQuantizer(disabled) +model.layers.25.self_attn.o_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.25.self_attn.q_bmm_quantizer TensorQuantizer(disabled) +model.layers.25.self_attn.k_bmm_quantizer TensorQuantizer((4, 3) bit fake per-tensor amax=4.48e+02(const) calibrator=MaxCalibrator quant) +model.layers.25.self_attn.v_bmm_quantizer TensorQuantizer((4, 3) bit fake per-tensor amax=4.48e+02(const) calibrator=MaxCalibrator quant) +model.layers.25.self_attn.p_bmm_quantizer TensorQuantizer(disabled) +model.layers.25.mlp.experts.gate_up_proj_input_quantizer TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.14e-01 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.gate_up_proj_weight_quantizers.0 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.gate_up_proj_weight_quantizers.1 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.86e-01 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.gate_up_proj_weight_quantizers.2 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.gate_up_proj_weight_quantizers.3 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.gate_up_proj_weight_quantizers.4 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.gate_up_proj_weight_quantizers.5 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.gate_up_proj_weight_quantizers.6 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.80e-01 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.gate_up_proj_weight_quantizers.7 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.gate_up_proj_weight_quantizers.8 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.gate_up_proj_weight_quantizers.9 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.gate_up_proj_weight_quantizers.10 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.gate_up_proj_weight_quantizers.11 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.gate_up_proj_weight_quantizers.12 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.82e-01 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.gate_up_proj_weight_quantizers.13 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.gate_up_proj_weight_quantizers.14 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.gate_up_proj_weight_quantizers.15 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.gate_up_proj_weight_quantizers.16 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.gate_up_proj_weight_quantizers.17 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.gate_up_proj_weight_quantizers.18 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.gate_up_proj_weight_quantizers.19 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.gate_up_proj_weight_quantizers.20 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.54e-01 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.gate_up_proj_weight_quantizers.21 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.13e-01 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.gate_up_proj_weight_quantizers.22 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.gate_up_proj_weight_quantizers.23 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.gate_up_proj_weight_quantizers.24 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.gate_up_proj_weight_quantizers.25 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.29e-01 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.gate_up_proj_weight_quantizers.26 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.gate_up_proj_weight_quantizers.27 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.gate_up_proj_weight_quantizers.28 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.gate_up_proj_weight_quantizers.29 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.gate_up_proj_weight_quantizers.30 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.gate_up_proj_weight_quantizers.31 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.gate_up_proj_weight_quantizers.32 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.gate_up_proj_weight_quantizers.33 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.gate_up_proj_weight_quantizers.34 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.gate_up_proj_weight_quantizers.35 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.gate_up_proj_weight_quantizers.36 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.gate_up_proj_weight_quantizers.37 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.gate_up_proj_weight_quantizers.38 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.gate_up_proj_weight_quantizers.39 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.gate_up_proj_weight_quantizers.40 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.gate_up_proj_weight_quantizers.41 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.gate_up_proj_weight_quantizers.42 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.gate_up_proj_weight_quantizers.43 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.gate_up_proj_weight_quantizers.44 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.gate_up_proj_weight_quantizers.45 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.gate_up_proj_weight_quantizers.46 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.gate_up_proj_weight_quantizers.47 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.gate_up_proj_weight_quantizers.48 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.28e-01 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.gate_up_proj_weight_quantizers.49 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.53e-01 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.gate_up_proj_weight_quantizers.50 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.gate_up_proj_weight_quantizers.51 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.36e-01 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.gate_up_proj_weight_quantizers.52 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.64e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.gate_up_proj_weight_quantizers.53 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.gate_up_proj_weight_quantizers.54 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.66e-01 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.gate_up_proj_weight_quantizers.55 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.gate_up_proj_weight_quantizers.56 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.gate_up_proj_weight_quantizers.57 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.88e-01 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.gate_up_proj_weight_quantizers.58 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.gate_up_proj_weight_quantizers.59 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.gate_up_proj_weight_quantizers.60 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.15e-01 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.gate_up_proj_weight_quantizers.61 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.gate_up_proj_weight_quantizers.62 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.gate_up_proj_weight_quantizers.63 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.gate_up_proj_weight_quantizers.64 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.71e-01 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.gate_up_proj_weight_quantizers.65 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.16e-01 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.gate_up_proj_weight_quantizers.66 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.gate_up_proj_weight_quantizers.67 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.gate_up_proj_weight_quantizers.68 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.gate_up_proj_weight_quantizers.69 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.gate_up_proj_weight_quantizers.70 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.gate_up_proj_weight_quantizers.71 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.gate_up_proj_weight_quantizers.72 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.gate_up_proj_weight_quantizers.73 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.gate_up_proj_weight_quantizers.74 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.gate_up_proj_weight_quantizers.75 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.gate_up_proj_weight_quantizers.76 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.gate_up_proj_weight_quantizers.77 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.gate_up_proj_weight_quantizers.78 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.gate_up_proj_weight_quantizers.79 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.20e-01 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.gate_up_proj_weight_quantizers.80 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.14e-01 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.gate_up_proj_weight_quantizers.81 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.gate_up_proj_weight_quantizers.82 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.gate_up_proj_weight_quantizers.83 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.21e-01 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.gate_up_proj_weight_quantizers.84 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.gate_up_proj_weight_quantizers.85 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.gate_up_proj_weight_quantizers.86 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.gate_up_proj_weight_quantizers.87 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.gate_up_proj_weight_quantizers.88 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.gate_up_proj_weight_quantizers.89 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.gate_up_proj_weight_quantizers.90 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.gate_up_proj_weight_quantizers.91 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.gate_up_proj_weight_quantizers.92 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.gate_up_proj_weight_quantizers.93 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.gate_up_proj_weight_quantizers.94 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.25e-01 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.gate_up_proj_weight_quantizers.95 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.gate_up_proj_weight_quantizers.96 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.gate_up_proj_weight_quantizers.97 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.18e-01 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.gate_up_proj_weight_quantizers.98 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.gate_up_proj_weight_quantizers.99 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.gate_up_proj_weight_quantizers.100 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.gate_up_proj_weight_quantizers.101 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.25e-01 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.gate_up_proj_weight_quantizers.102 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.gate_up_proj_weight_quantizers.103 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.gate_up_proj_weight_quantizers.104 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.gate_up_proj_weight_quantizers.105 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.gate_up_proj_weight_quantizers.106 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.gate_up_proj_weight_quantizers.107 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.gate_up_proj_weight_quantizers.108 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.gate_up_proj_weight_quantizers.109 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.gate_up_proj_weight_quantizers.110 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.gate_up_proj_weight_quantizers.111 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.01e-01 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.gate_up_proj_weight_quantizers.112 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.gate_up_proj_weight_quantizers.113 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.41e-01 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.gate_up_proj_weight_quantizers.114 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.gate_up_proj_weight_quantizers.115 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.18e-01 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.gate_up_proj_weight_quantizers.116 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.gate_up_proj_weight_quantizers.117 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.gate_up_proj_weight_quantizers.118 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.gate_up_proj_weight_quantizers.119 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.gate_up_proj_weight_quantizers.120 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.gate_up_proj_weight_quantizers.121 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.gate_up_proj_weight_quantizers.122 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.31e-01 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.gate_up_proj_weight_quantizers.123 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.gate_up_proj_weight_quantizers.124 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.gate_up_proj_weight_quantizers.125 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.gate_up_proj_weight_quantizers.126 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.gate_up_proj_weight_quantizers.127 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.gate_up_proj_weight_quantizers.128 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.15e-01 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.gate_up_proj_weight_quantizers.129 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.gate_up_proj_weight_quantizers.130 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.gate_up_proj_weight_quantizers.131 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.gate_up_proj_weight_quantizers.132 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.gate_up_proj_weight_quantizers.133 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.46e-01 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.gate_up_proj_weight_quantizers.134 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.gate_up_proj_weight_quantizers.135 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.gate_up_proj_weight_quantizers.136 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.gate_up_proj_weight_quantizers.137 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.gate_up_proj_weight_quantizers.138 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.gate_up_proj_weight_quantizers.139 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.gate_up_proj_weight_quantizers.140 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.gate_up_proj_weight_quantizers.141 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.19e-01 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.gate_up_proj_weight_quantizers.142 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.gate_up_proj_weight_quantizers.143 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.gate_up_proj_weight_quantizers.144 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.18e-01 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.gate_up_proj_weight_quantizers.145 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.gate_up_proj_weight_quantizers.146 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.gate_up_proj_weight_quantizers.147 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.gate_up_proj_weight_quantizers.148 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.gate_up_proj_weight_quantizers.149 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.gate_up_proj_weight_quantizers.150 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.gate_up_proj_weight_quantizers.151 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.16e-01 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.gate_up_proj_weight_quantizers.152 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.gate_up_proj_weight_quantizers.153 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.gate_up_proj_weight_quantizers.154 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.gate_up_proj_weight_quantizers.155 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.gate_up_proj_weight_quantizers.156 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.gate_up_proj_weight_quantizers.157 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.gate_up_proj_weight_quantizers.158 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.gate_up_proj_weight_quantizers.159 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.gate_up_proj_weight_quantizers.160 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.gate_up_proj_weight_quantizers.161 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.gate_up_proj_weight_quantizers.162 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.gate_up_proj_weight_quantizers.163 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.gate_up_proj_weight_quantizers.164 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.gate_up_proj_weight_quantizers.165 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.24e-01 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.gate_up_proj_weight_quantizers.166 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.gate_up_proj_weight_quantizers.167 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.gate_up_proj_weight_quantizers.168 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.68e-01 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.gate_up_proj_weight_quantizers.169 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.41e-01 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.gate_up_proj_weight_quantizers.170 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.gate_up_proj_weight_quantizers.171 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.gate_up_proj_weight_quantizers.172 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.gate_up_proj_weight_quantizers.173 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.gate_up_proj_weight_quantizers.174 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.gate_up_proj_weight_quantizers.175 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.gate_up_proj_weight_quantizers.176 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.24e-01 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.gate_up_proj_weight_quantizers.177 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.gate_up_proj_weight_quantizers.178 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.59e-01 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.gate_up_proj_weight_quantizers.179 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.gate_up_proj_weight_quantizers.180 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.gate_up_proj_weight_quantizers.181 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.gate_up_proj_weight_quantizers.182 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.gate_up_proj_weight_quantizers.183 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.24e-01 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.gate_up_proj_weight_quantizers.184 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.36e-01 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.gate_up_proj_weight_quantizers.185 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.gate_up_proj_weight_quantizers.186 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.gate_up_proj_weight_quantizers.187 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.gate_up_proj_weight_quantizers.188 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.gate_up_proj_weight_quantizers.189 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.gate_up_proj_weight_quantizers.190 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.gate_up_proj_weight_quantizers.191 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.gate_up_proj_weight_quantizers.192 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.23e-01 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.gate_up_proj_weight_quantizers.193 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.gate_up_proj_weight_quantizers.194 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.gate_up_proj_weight_quantizers.195 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.11e-01 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.gate_up_proj_weight_quantizers.196 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.gate_up_proj_weight_quantizers.197 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.gate_up_proj_weight_quantizers.198 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.41e-01 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.gate_up_proj_weight_quantizers.199 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.19e-01 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.gate_up_proj_weight_quantizers.200 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.gate_up_proj_weight_quantizers.201 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.gate_up_proj_weight_quantizers.202 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.gate_up_proj_weight_quantizers.203 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.gate_up_proj_weight_quantizers.204 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.gate_up_proj_weight_quantizers.205 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.gate_up_proj_weight_quantizers.206 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.gate_up_proj_weight_quantizers.207 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.gate_up_proj_weight_quantizers.208 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.15e-01 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.gate_up_proj_weight_quantizers.209 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.gate_up_proj_weight_quantizers.210 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.gate_up_proj_weight_quantizers.211 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.gate_up_proj_weight_quantizers.212 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.gate_up_proj_weight_quantizers.213 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.gate_up_proj_weight_quantizers.214 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.gate_up_proj_weight_quantizers.215 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.16e-01 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.gate_up_proj_weight_quantizers.216 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.gate_up_proj_weight_quantizers.217 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.gate_up_proj_weight_quantizers.218 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.gate_up_proj_weight_quantizers.219 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.gate_up_proj_weight_quantizers.220 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.gate_up_proj_weight_quantizers.221 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.gate_up_proj_weight_quantizers.222 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.gate_up_proj_weight_quantizers.223 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.gate_up_proj_weight_quantizers.224 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.gate_up_proj_weight_quantizers.225 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.gate_up_proj_weight_quantizers.226 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.gate_up_proj_weight_quantizers.227 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.gate_up_proj_weight_quantizers.228 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.gate_up_proj_weight_quantizers.229 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.gate_up_proj_weight_quantizers.230 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.gate_up_proj_weight_quantizers.231 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.gate_up_proj_weight_quantizers.232 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.gate_up_proj_weight_quantizers.233 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.gate_up_proj_weight_quantizers.234 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.gate_up_proj_weight_quantizers.235 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.gate_up_proj_weight_quantizers.236 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.gate_up_proj_weight_quantizers.237 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.gate_up_proj_weight_quantizers.238 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.gate_up_proj_weight_quantizers.239 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.gate_up_proj_weight_quantizers.240 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.gate_up_proj_weight_quantizers.241 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.gate_up_proj_weight_quantizers.242 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.gate_up_proj_weight_quantizers.243 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.40e-01 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.gate_up_proj_weight_quantizers.244 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.gate_up_proj_weight_quantizers.245 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.94e-01 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.gate_up_proj_weight_quantizers.246 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.gate_up_proj_weight_quantizers.247 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.gate_up_proj_weight_quantizers.248 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.15e-01 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.gate_up_proj_weight_quantizers.249 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.gate_up_proj_weight_quantizers.250 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.gate_up_proj_weight_quantizers.251 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.gate_up_proj_weight_quantizers.252 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.gate_up_proj_weight_quantizers.253 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.gate_up_proj_weight_quantizers.254 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.gate_up_proj_weight_quantizers.255 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.down_proj_input_quantizer TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=6.62e+00 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.down_proj_weight_quantizers.0 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.down_proj_weight_quantizers.1 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.18e-01 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.down_proj_weight_quantizers.2 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.down_proj_weight_quantizers.3 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.56e-01 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.down_proj_weight_quantizers.4 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.down_proj_weight_quantizers.5 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.down_proj_weight_quantizers.6 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.52e-01 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.down_proj_weight_quantizers.7 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.down_proj_weight_quantizers.8 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.down_proj_weight_quantizers.9 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.down_proj_weight_quantizers.10 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.down_proj_weight_quantizers.11 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.down_proj_weight_quantizers.12 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.48e-01 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.down_proj_weight_quantizers.13 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.45e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.down_proj_weight_quantizers.14 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.64e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.down_proj_weight_quantizers.15 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.down_proj_weight_quantizers.16 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.down_proj_weight_quantizers.17 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.down_proj_weight_quantizers.18 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.down_proj_weight_quantizers.19 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.down_proj_weight_quantizers.20 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.43e-01 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.down_proj_weight_quantizers.21 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.down_proj_weight_quantizers.22 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.down_proj_weight_quantizers.23 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.down_proj_weight_quantizers.24 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.down_proj_weight_quantizers.25 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.32e-01 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.down_proj_weight_quantizers.26 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.down_proj_weight_quantizers.27 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.down_proj_weight_quantizers.28 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.down_proj_weight_quantizers.29 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.down_proj_weight_quantizers.30 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.down_proj_weight_quantizers.31 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.down_proj_weight_quantizers.32 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.down_proj_weight_quantizers.33 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.down_proj_weight_quantizers.34 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.62e-01 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.down_proj_weight_quantizers.35 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.down_proj_weight_quantizers.36 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.64e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.down_proj_weight_quantizers.37 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.down_proj_weight_quantizers.38 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.down_proj_weight_quantizers.39 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.23e-01 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.down_proj_weight_quantizers.40 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.down_proj_weight_quantizers.41 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.50e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.down_proj_weight_quantizers.42 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.down_proj_weight_quantizers.43 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.down_proj_weight_quantizers.44 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.down_proj_weight_quantizers.45 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.down_proj_weight_quantizers.46 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.down_proj_weight_quantizers.47 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.down_proj_weight_quantizers.48 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.46e-01 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.down_proj_weight_quantizers.49 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.down_proj_weight_quantizers.50 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.down_proj_weight_quantizers.51 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.down_proj_weight_quantizers.52 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.down_proj_weight_quantizers.53 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.down_proj_weight_quantizers.54 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.36e-01 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.down_proj_weight_quantizers.55 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.down_proj_weight_quantizers.56 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.down_proj_weight_quantizers.57 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.down_proj_weight_quantizers.58 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.down_proj_weight_quantizers.59 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.down_proj_weight_quantizers.60 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.53e-01 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.down_proj_weight_quantizers.61 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.down_proj_weight_quantizers.62 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.32e-01 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.down_proj_weight_quantizers.63 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.down_proj_weight_quantizers.64 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.down_proj_weight_quantizers.65 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.down_proj_weight_quantizers.66 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.down_proj_weight_quantizers.67 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.down_proj_weight_quantizers.68 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.down_proj_weight_quantizers.69 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.64e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.down_proj_weight_quantizers.70 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.down_proj_weight_quantizers.71 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.down_proj_weight_quantizers.72 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.down_proj_weight_quantizers.73 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.down_proj_weight_quantizers.74 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.down_proj_weight_quantizers.75 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.down_proj_weight_quantizers.76 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.down_proj_weight_quantizers.77 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.59e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.down_proj_weight_quantizers.78 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.down_proj_weight_quantizers.79 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.down_proj_weight_quantizers.80 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.35e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.down_proj_weight_quantizers.81 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.down_proj_weight_quantizers.82 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.down_proj_weight_quantizers.83 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.34e-01 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.down_proj_weight_quantizers.84 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.59e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.down_proj_weight_quantizers.85 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.down_proj_weight_quantizers.86 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.down_proj_weight_quantizers.87 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.54e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.down_proj_weight_quantizers.88 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.down_proj_weight_quantizers.89 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.down_proj_weight_quantizers.90 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.down_proj_weight_quantizers.91 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.45e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.down_proj_weight_quantizers.92 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.down_proj_weight_quantizers.93 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.down_proj_weight_quantizers.94 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.18e-01 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.down_proj_weight_quantizers.95 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.down_proj_weight_quantizers.96 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.down_proj_weight_quantizers.97 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.31e-01 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.down_proj_weight_quantizers.98 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.down_proj_weight_quantizers.99 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.down_proj_weight_quantizers.100 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.down_proj_weight_quantizers.101 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.down_proj_weight_quantizers.102 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.down_proj_weight_quantizers.103 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.down_proj_weight_quantizers.104 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.down_proj_weight_quantizers.105 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.down_proj_weight_quantizers.106 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.down_proj_weight_quantizers.107 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.down_proj_weight_quantizers.108 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.down_proj_weight_quantizers.109 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.down_proj_weight_quantizers.110 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.down_proj_weight_quantizers.111 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.80e-01 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.down_proj_weight_quantizers.112 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.down_proj_weight_quantizers.113 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.33e-01 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.down_proj_weight_quantizers.114 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.64e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.down_proj_weight_quantizers.115 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.down_proj_weight_quantizers.116 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.down_proj_weight_quantizers.117 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.down_proj_weight_quantizers.118 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.down_proj_weight_quantizers.119 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.down_proj_weight_quantizers.120 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.down_proj_weight_quantizers.121 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.down_proj_weight_quantizers.122 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.down_proj_weight_quantizers.123 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.down_proj_weight_quantizers.124 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.40e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.down_proj_weight_quantizers.125 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.down_proj_weight_quantizers.126 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.down_proj_weight_quantizers.127 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.down_proj_weight_quantizers.128 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.36e-01 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.down_proj_weight_quantizers.129 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.down_proj_weight_quantizers.130 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.down_proj_weight_quantizers.131 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.down_proj_weight_quantizers.132 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.64e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.down_proj_weight_quantizers.133 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.down_proj_weight_quantizers.134 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.50e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.down_proj_weight_quantizers.135 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.down_proj_weight_quantizers.136 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.59e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.down_proj_weight_quantizers.137 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.down_proj_weight_quantizers.138 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.down_proj_weight_quantizers.139 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.14e-01 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.down_proj_weight_quantizers.140 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.down_proj_weight_quantizers.141 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.down_proj_weight_quantizers.142 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.24e-01 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.down_proj_weight_quantizers.143 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.down_proj_weight_quantizers.144 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.36e-01 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.down_proj_weight_quantizers.145 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.down_proj_weight_quantizers.146 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.24e-01 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.down_proj_weight_quantizers.147 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.down_proj_weight_quantizers.148 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.64e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.down_proj_weight_quantizers.149 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.down_proj_weight_quantizers.150 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.down_proj_weight_quantizers.151 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.down_proj_weight_quantizers.152 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.down_proj_weight_quantizers.153 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.down_proj_weight_quantizers.154 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.54e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.down_proj_weight_quantizers.155 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.64e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.down_proj_weight_quantizers.156 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.down_proj_weight_quantizers.157 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.down_proj_weight_quantizers.158 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.down_proj_weight_quantizers.159 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.50e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.down_proj_weight_quantizers.160 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.down_proj_weight_quantizers.161 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.down_proj_weight_quantizers.162 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.40e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.down_proj_weight_quantizers.163 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.down_proj_weight_quantizers.164 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.down_proj_weight_quantizers.165 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.down_proj_weight_quantizers.166 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.down_proj_weight_quantizers.167 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.down_proj_weight_quantizers.168 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=5.16e-01 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.down_proj_weight_quantizers.169 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.29e-01 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.down_proj_weight_quantizers.170 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.down_proj_weight_quantizers.171 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.down_proj_weight_quantizers.172 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.down_proj_weight_quantizers.173 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.50e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.down_proj_weight_quantizers.174 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.down_proj_weight_quantizers.175 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.57e-01 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.down_proj_weight_quantizers.176 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.down_proj_weight_quantizers.177 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.down_proj_weight_quantizers.178 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.32e-01 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.down_proj_weight_quantizers.179 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.down_proj_weight_quantizers.180 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.down_proj_weight_quantizers.181 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.down_proj_weight_quantizers.182 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.down_proj_weight_quantizers.183 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.down_proj_weight_quantizers.184 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.49e-01 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.down_proj_weight_quantizers.185 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.down_proj_weight_quantizers.186 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.down_proj_weight_quantizers.187 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.down_proj_weight_quantizers.188 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.down_proj_weight_quantizers.189 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.down_proj_weight_quantizers.190 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.down_proj_weight_quantizers.191 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.down_proj_weight_quantizers.192 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.49e-01 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.down_proj_weight_quantizers.193 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.down_proj_weight_quantizers.194 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.down_proj_weight_quantizers.195 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.44e-01 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.down_proj_weight_quantizers.196 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.down_proj_weight_quantizers.197 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.down_proj_weight_quantizers.198 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.28e-01 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.down_proj_weight_quantizers.199 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.down_proj_weight_quantizers.200 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.down_proj_weight_quantizers.201 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.down_proj_weight_quantizers.202 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.down_proj_weight_quantizers.203 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.down_proj_weight_quantizers.204 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.down_proj_weight_quantizers.205 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.55e-01 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.down_proj_weight_quantizers.206 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.down_proj_weight_quantizers.207 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.down_proj_weight_quantizers.208 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.down_proj_weight_quantizers.209 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.down_proj_weight_quantizers.210 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.down_proj_weight_quantizers.211 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.down_proj_weight_quantizers.212 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.down_proj_weight_quantizers.213 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.64e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.down_proj_weight_quantizers.214 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.down_proj_weight_quantizers.215 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.down_proj_weight_quantizers.216 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.down_proj_weight_quantizers.217 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.30e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.down_proj_weight_quantizers.218 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.down_proj_weight_quantizers.219 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.down_proj_weight_quantizers.220 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.down_proj_weight_quantizers.221 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.40e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.down_proj_weight_quantizers.222 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.27e-01 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.down_proj_weight_quantizers.223 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.down_proj_weight_quantizers.224 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.down_proj_weight_quantizers.225 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.45e-01 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.down_proj_weight_quantizers.226 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.59e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.down_proj_weight_quantizers.227 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.down_proj_weight_quantizers.228 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.down_proj_weight_quantizers.229 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.54e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.down_proj_weight_quantizers.230 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.59e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.down_proj_weight_quantizers.231 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.down_proj_weight_quantizers.232 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.down_proj_weight_quantizers.233 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.24e-01 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.down_proj_weight_quantizers.234 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.down_proj_weight_quantizers.235 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.down_proj_weight_quantizers.236 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.down_proj_weight_quantizers.237 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.down_proj_weight_quantizers.238 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.down_proj_weight_quantizers.239 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.down_proj_weight_quantizers.240 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.down_proj_weight_quantizers.241 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.down_proj_weight_quantizers.242 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.59e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.down_proj_weight_quantizers.243 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.61e-01 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.down_proj_weight_quantizers.244 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.64e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.down_proj_weight_quantizers.245 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.58e-01 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.down_proj_weight_quantizers.246 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.59e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.down_proj_weight_quantizers.247 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.down_proj_weight_quantizers.248 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.down_proj_weight_quantizers.249 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.down_proj_weight_quantizers.250 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.down_proj_weight_quantizers.251 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.59e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.down_proj_weight_quantizers.252 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.down_proj_weight_quantizers.253 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.down_proj_weight_quantizers.254 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.40e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.down_proj_weight_quantizers.255 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.shared_experts.gate_proj.input_quantizer TensorQuantizer(disabled) +model.layers.25.mlp.shared_experts.gate_proj.output_quantizer TensorQuantizer(disabled) +model.layers.25.mlp.shared_experts.gate_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.25.mlp.shared_experts.up_proj.input_quantizer TensorQuantizer(disabled) +model.layers.25.mlp.shared_experts.up_proj.output_quantizer TensorQuantizer(disabled) +model.layers.25.mlp.shared_experts.up_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.25.mlp.shared_experts.down_proj.input_quantizer TensorQuantizer(disabled) +model.layers.25.mlp.shared_experts.down_proj.output_quantizer TensorQuantizer(disabled) +model.layers.25.mlp.shared_experts.down_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.26.self_attn.q_a_proj.input_quantizer TensorQuantizer(disabled) +model.layers.26.self_attn.q_a_proj.output_quantizer TensorQuantizer(disabled) +model.layers.26.self_attn.q_a_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.26.self_attn.q_b_proj.input_quantizer TensorQuantizer(disabled) +model.layers.26.self_attn.q_b_proj.output_quantizer TensorQuantizer(disabled) +model.layers.26.self_attn.q_b_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.26.self_attn.kv_a_proj_with_mqa.input_quantizer TensorQuantizer(disabled) +model.layers.26.self_attn.kv_a_proj_with_mqa.output_quantizer TensorQuantizer(disabled) +model.layers.26.self_attn.kv_a_proj_with_mqa.weight_quantizer TensorQuantizer(disabled) +model.layers.26.self_attn.kv_b_proj.input_quantizer TensorQuantizer(disabled) +model.layers.26.self_attn.kv_b_proj.output_quantizer TensorQuantizer(disabled) +model.layers.26.self_attn.kv_b_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.26.self_attn.o_proj.input_quantizer TensorQuantizer(disabled) +model.layers.26.self_attn.o_proj.output_quantizer TensorQuantizer(disabled) +model.layers.26.self_attn.o_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.26.self_attn.indexer.wq_b.input_quantizer TensorQuantizer(disabled) +model.layers.26.self_attn.indexer.wq_b.output_quantizer TensorQuantizer(disabled) +model.layers.26.self_attn.indexer.wq_b.weight_quantizer TensorQuantizer(disabled) +model.layers.26.self_attn.indexer.wk.input_quantizer TensorQuantizer(disabled) +model.layers.26.self_attn.indexer.wk.output_quantizer TensorQuantizer(disabled) +model.layers.26.self_attn.indexer.wk.weight_quantizer TensorQuantizer(disabled) +model.layers.26.self_attn.indexer.k_norm.input_quantizer TensorQuantizer(disabled) +model.layers.26.self_attn.indexer.k_norm.output_quantizer TensorQuantizer(disabled) +model.layers.26.self_attn.indexer.weights_proj.input_quantizer TensorQuantizer(disabled) +model.layers.26.self_attn.indexer.weights_proj.output_quantizer TensorQuantizer(disabled) +model.layers.26.self_attn.indexer.weights_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.26.self_attn.q_bmm_quantizer TensorQuantizer(disabled) +model.layers.26.self_attn.k_bmm_quantizer TensorQuantizer((4, 3) bit fake per-tensor amax=4.48e+02(const) calibrator=MaxCalibrator quant) +model.layers.26.self_attn.v_bmm_quantizer TensorQuantizer((4, 3) bit fake per-tensor amax=4.48e+02(const) calibrator=MaxCalibrator quant) +model.layers.26.self_attn.p_bmm_quantizer TensorQuantizer(disabled) +model.layers.26.mlp.experts.gate_up_proj_input_quantizer TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.91e-01 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.gate_up_proj_weight_quantizers.0 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.gate_up_proj_weight_quantizers.1 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.gate_up_proj_weight_quantizers.2 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.gate_up_proj_weight_quantizers.3 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.54e-01 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.gate_up_proj_weight_quantizers.4 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.gate_up_proj_weight_quantizers.5 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.gate_up_proj_weight_quantizers.6 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.gate_up_proj_weight_quantizers.7 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.44e-01 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.gate_up_proj_weight_quantizers.8 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.17e-01 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.gate_up_proj_weight_quantizers.9 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.gate_up_proj_weight_quantizers.10 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.gate_up_proj_weight_quantizers.11 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.38e-01 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.gate_up_proj_weight_quantizers.12 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.gate_up_proj_weight_quantizers.13 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.gate_up_proj_weight_quantizers.14 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.gate_up_proj_weight_quantizers.15 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.gate_up_proj_weight_quantizers.16 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.gate_up_proj_weight_quantizers.17 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.gate_up_proj_weight_quantizers.18 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.gate_up_proj_weight_quantizers.19 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.gate_up_proj_weight_quantizers.20 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.gate_up_proj_weight_quantizers.21 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.gate_up_proj_weight_quantizers.22 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.gate_up_proj_weight_quantizers.23 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.gate_up_proj_weight_quantizers.24 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.gate_up_proj_weight_quantizers.25 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.gate_up_proj_weight_quantizers.26 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.gate_up_proj_weight_quantizers.27 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.gate_up_proj_weight_quantizers.28 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.gate_up_proj_weight_quantizers.29 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.gate_up_proj_weight_quantizers.30 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.gate_up_proj_weight_quantizers.31 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.gate_up_proj_weight_quantizers.32 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.37e-01 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.gate_up_proj_weight_quantizers.33 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.gate_up_proj_weight_quantizers.34 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.gate_up_proj_weight_quantizers.35 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.gate_up_proj_weight_quantizers.36 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.56e-01 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.gate_up_proj_weight_quantizers.37 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.gate_up_proj_weight_quantizers.38 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.gate_up_proj_weight_quantizers.39 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.14e-01 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.gate_up_proj_weight_quantizers.40 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.32e-01 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.gate_up_proj_weight_quantizers.41 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.gate_up_proj_weight_quantizers.42 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.gate_up_proj_weight_quantizers.43 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.gate_up_proj_weight_quantizers.44 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.gate_up_proj_weight_quantizers.45 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.gate_up_proj_weight_quantizers.46 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.39e-01 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.gate_up_proj_weight_quantizers.47 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.gate_up_proj_weight_quantizers.48 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.gate_up_proj_weight_quantizers.49 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.gate_up_proj_weight_quantizers.50 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.gate_up_proj_weight_quantizers.51 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.gate_up_proj_weight_quantizers.52 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.gate_up_proj_weight_quantizers.53 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.gate_up_proj_weight_quantizers.54 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.gate_up_proj_weight_quantizers.55 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.gate_up_proj_weight_quantizers.56 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.gate_up_proj_weight_quantizers.57 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.gate_up_proj_weight_quantizers.58 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.gate_up_proj_weight_quantizers.59 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.gate_up_proj_weight_quantizers.60 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.gate_up_proj_weight_quantizers.61 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.gate_up_proj_weight_quantizers.62 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.gate_up_proj_weight_quantizers.63 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.gate_up_proj_weight_quantizers.64 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.gate_up_proj_weight_quantizers.65 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.gate_up_proj_weight_quantizers.66 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.gate_up_proj_weight_quantizers.67 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.gate_up_proj_weight_quantizers.68 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.gate_up_proj_weight_quantizers.69 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.gate_up_proj_weight_quantizers.70 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.90e-01 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.gate_up_proj_weight_quantizers.71 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.gate_up_proj_weight_quantizers.72 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.gate_up_proj_weight_quantizers.73 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.gate_up_proj_weight_quantizers.74 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.gate_up_proj_weight_quantizers.75 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.44e-01 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.gate_up_proj_weight_quantizers.76 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.47e-01 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.gate_up_proj_weight_quantizers.77 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.gate_up_proj_weight_quantizers.78 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.gate_up_proj_weight_quantizers.79 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.gate_up_proj_weight_quantizers.80 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.24e-01 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.gate_up_proj_weight_quantizers.81 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.gate_up_proj_weight_quantizers.82 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.gate_up_proj_weight_quantizers.83 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.gate_up_proj_weight_quantizers.84 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.gate_up_proj_weight_quantizers.85 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.gate_up_proj_weight_quantizers.86 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.gate_up_proj_weight_quantizers.87 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.37e-01 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.gate_up_proj_weight_quantizers.88 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.gate_up_proj_weight_quantizers.89 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.gate_up_proj_weight_quantizers.90 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.gate_up_proj_weight_quantizers.91 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.15e-01 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.gate_up_proj_weight_quantizers.92 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.gate_up_proj_weight_quantizers.93 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.gate_up_proj_weight_quantizers.94 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.gate_up_proj_weight_quantizers.95 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.gate_up_proj_weight_quantizers.96 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.gate_up_proj_weight_quantizers.97 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.gate_up_proj_weight_quantizers.98 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.56e-01 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.gate_up_proj_weight_quantizers.99 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.gate_up_proj_weight_quantizers.100 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.gate_up_proj_weight_quantizers.101 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.gate_up_proj_weight_quantizers.102 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.gate_up_proj_weight_quantizers.103 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.gate_up_proj_weight_quantizers.104 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.gate_up_proj_weight_quantizers.105 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.57e-01 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.gate_up_proj_weight_quantizers.106 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.17e-01 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.gate_up_proj_weight_quantizers.107 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.gate_up_proj_weight_quantizers.108 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.gate_up_proj_weight_quantizers.109 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.gate_up_proj_weight_quantizers.110 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.gate_up_proj_weight_quantizers.111 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.gate_up_proj_weight_quantizers.112 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.gate_up_proj_weight_quantizers.113 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.gate_up_proj_weight_quantizers.114 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.gate_up_proj_weight_quantizers.115 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.gate_up_proj_weight_quantizers.116 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.gate_up_proj_weight_quantizers.117 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.gate_up_proj_weight_quantizers.118 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.gate_up_proj_weight_quantizers.119 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.61e-01 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.gate_up_proj_weight_quantizers.120 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.gate_up_proj_weight_quantizers.121 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.13e-01 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.gate_up_proj_weight_quantizers.122 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.gate_up_proj_weight_quantizers.123 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.gate_up_proj_weight_quantizers.124 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.gate_up_proj_weight_quantizers.125 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.gate_up_proj_weight_quantizers.126 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.gate_up_proj_weight_quantizers.127 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.gate_up_proj_weight_quantizers.128 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.gate_up_proj_weight_quantizers.129 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.gate_up_proj_weight_quantizers.130 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.gate_up_proj_weight_quantizers.131 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.64e-01 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.gate_up_proj_weight_quantizers.132 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.gate_up_proj_weight_quantizers.133 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.35e-01 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.gate_up_proj_weight_quantizers.134 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.gate_up_proj_weight_quantizers.135 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.gate_up_proj_weight_quantizers.136 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.21e-01 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.gate_up_proj_weight_quantizers.137 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.gate_up_proj_weight_quantizers.138 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.gate_up_proj_weight_quantizers.139 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.gate_up_proj_weight_quantizers.140 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.gate_up_proj_weight_quantizers.141 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.gate_up_proj_weight_quantizers.142 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.21e-01 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.gate_up_proj_weight_quantizers.143 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.gate_up_proj_weight_quantizers.144 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.gate_up_proj_weight_quantizers.145 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.gate_up_proj_weight_quantizers.146 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.gate_up_proj_weight_quantizers.147 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.gate_up_proj_weight_quantizers.148 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.gate_up_proj_weight_quantizers.149 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.gate_up_proj_weight_quantizers.150 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.gate_up_proj_weight_quantizers.151 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.gate_up_proj_weight_quantizers.152 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.gate_up_proj_weight_quantizers.153 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.26e-01 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.gate_up_proj_weight_quantizers.154 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.gate_up_proj_weight_quantizers.155 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.88e-01 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.gate_up_proj_weight_quantizers.156 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.gate_up_proj_weight_quantizers.157 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.gate_up_proj_weight_quantizers.158 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.gate_up_proj_weight_quantizers.159 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.gate_up_proj_weight_quantizers.160 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.gate_up_proj_weight_quantizers.161 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.gate_up_proj_weight_quantizers.162 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.32e-01 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.gate_up_proj_weight_quantizers.163 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.gate_up_proj_weight_quantizers.164 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.gate_up_proj_weight_quantizers.165 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.gate_up_proj_weight_quantizers.166 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.07e-01 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.gate_up_proj_weight_quantizers.167 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.gate_up_proj_weight_quantizers.168 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.gate_up_proj_weight_quantizers.169 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.gate_up_proj_weight_quantizers.170 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.gate_up_proj_weight_quantizers.171 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.gate_up_proj_weight_quantizers.172 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.39e-01 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.gate_up_proj_weight_quantizers.173 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.gate_up_proj_weight_quantizers.174 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.gate_up_proj_weight_quantizers.175 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.gate_up_proj_weight_quantizers.176 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.gate_up_proj_weight_quantizers.177 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.gate_up_proj_weight_quantizers.178 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.gate_up_proj_weight_quantizers.179 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.gate_up_proj_weight_quantizers.180 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.gate_up_proj_weight_quantizers.181 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.gate_up_proj_weight_quantizers.182 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.37e-01 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.gate_up_proj_weight_quantizers.183 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.gate_up_proj_weight_quantizers.184 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.gate_up_proj_weight_quantizers.185 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.gate_up_proj_weight_quantizers.186 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.gate_up_proj_weight_quantizers.187 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.gate_up_proj_weight_quantizers.188 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.gate_up_proj_weight_quantizers.189 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.gate_up_proj_weight_quantizers.190 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.gate_up_proj_weight_quantizers.191 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.gate_up_proj_weight_quantizers.192 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.gate_up_proj_weight_quantizers.193 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.gate_up_proj_weight_quantizers.194 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.37e-01 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.gate_up_proj_weight_quantizers.195 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.gate_up_proj_weight_quantizers.196 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.gate_up_proj_weight_quantizers.197 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.gate_up_proj_weight_quantizers.198 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.gate_up_proj_weight_quantizers.199 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.gate_up_proj_weight_quantizers.200 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.16e-01 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.gate_up_proj_weight_quantizers.201 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.gate_up_proj_weight_quantizers.202 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.gate_up_proj_weight_quantizers.203 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.gate_up_proj_weight_quantizers.204 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.gate_up_proj_weight_quantizers.205 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.gate_up_proj_weight_quantizers.206 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.gate_up_proj_weight_quantizers.207 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.41e-01 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.gate_up_proj_weight_quantizers.208 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.gate_up_proj_weight_quantizers.209 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.gate_up_proj_weight_quantizers.210 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.gate_up_proj_weight_quantizers.211 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.gate_up_proj_weight_quantizers.212 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.gate_up_proj_weight_quantizers.213 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.gate_up_proj_weight_quantizers.214 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.gate_up_proj_weight_quantizers.215 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.gate_up_proj_weight_quantizers.216 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.gate_up_proj_weight_quantizers.217 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.gate_up_proj_weight_quantizers.218 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.gate_up_proj_weight_quantizers.219 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.gate_up_proj_weight_quantizers.220 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.gate_up_proj_weight_quantizers.221 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.gate_up_proj_weight_quantizers.222 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.gate_up_proj_weight_quantizers.223 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.gate_up_proj_weight_quantizers.224 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.gate_up_proj_weight_quantizers.225 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.gate_up_proj_weight_quantizers.226 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.gate_up_proj_weight_quantizers.227 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.gate_up_proj_weight_quantizers.228 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.gate_up_proj_weight_quantizers.229 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.gate_up_proj_weight_quantizers.230 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.gate_up_proj_weight_quantizers.231 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.gate_up_proj_weight_quantizers.232 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.gate_up_proj_weight_quantizers.233 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.gate_up_proj_weight_quantizers.234 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.gate_up_proj_weight_quantizers.235 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.gate_up_proj_weight_quantizers.236 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.gate_up_proj_weight_quantizers.237 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.gate_up_proj_weight_quantizers.238 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.gate_up_proj_weight_quantizers.239 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.gate_up_proj_weight_quantizers.240 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.gate_up_proj_weight_quantizers.241 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.15e-01 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.gate_up_proj_weight_quantizers.242 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.gate_up_proj_weight_quantizers.243 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.gate_up_proj_weight_quantizers.244 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.gate_up_proj_weight_quantizers.245 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.gate_up_proj_weight_quantizers.246 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.20e-01 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.gate_up_proj_weight_quantizers.247 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.gate_up_proj_weight_quantizers.248 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.gate_up_proj_weight_quantizers.249 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.78e-01 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.gate_up_proj_weight_quantizers.250 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.gate_up_proj_weight_quantizers.251 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.gate_up_proj_weight_quantizers.252 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.gate_up_proj_weight_quantizers.253 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.46e-01 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.gate_up_proj_weight_quantizers.254 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.gate_up_proj_weight_quantizers.255 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.down_proj_input_quantizer TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.38e+00 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.down_proj_weight_quantizers.0 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.down_proj_weight_quantizers.1 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.down_proj_weight_quantizers.2 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.down_proj_weight_quantizers.3 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.59e-01 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.down_proj_weight_quantizers.4 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.down_proj_weight_quantizers.5 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.down_proj_weight_quantizers.6 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.down_proj_weight_quantizers.7 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.04e-01 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.down_proj_weight_quantizers.8 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.down_proj_weight_quantizers.9 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.down_proj_weight_quantizers.10 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.down_proj_weight_quantizers.11 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.down_proj_weight_quantizers.12 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.down_proj_weight_quantizers.13 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.down_proj_weight_quantizers.14 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.down_proj_weight_quantizers.15 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.down_proj_weight_quantizers.16 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.down_proj_weight_quantizers.17 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.down_proj_weight_quantizers.18 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.down_proj_weight_quantizers.19 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.down_proj_weight_quantizers.20 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.down_proj_weight_quantizers.21 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.down_proj_weight_quantizers.22 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.down_proj_weight_quantizers.23 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.30e-01 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.down_proj_weight_quantizers.24 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.64e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.down_proj_weight_quantizers.25 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.down_proj_weight_quantizers.26 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.down_proj_weight_quantizers.27 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.down_proj_weight_quantizers.28 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.31e-01 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.down_proj_weight_quantizers.29 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.down_proj_weight_quantizers.30 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.down_proj_weight_quantizers.31 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.54e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.down_proj_weight_quantizers.32 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.down_proj_weight_quantizers.33 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.down_proj_weight_quantizers.34 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.down_proj_weight_quantizers.35 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.down_proj_weight_quantizers.36 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.54e-01 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.down_proj_weight_quantizers.37 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.64e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.down_proj_weight_quantizers.38 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.45e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.down_proj_weight_quantizers.39 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.down_proj_weight_quantizers.40 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.down_proj_weight_quantizers.41 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.down_proj_weight_quantizers.42 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.down_proj_weight_quantizers.43 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.down_proj_weight_quantizers.44 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.down_proj_weight_quantizers.45 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.down_proj_weight_quantizers.46 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.down_proj_weight_quantizers.47 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.20e-01 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.down_proj_weight_quantizers.48 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.down_proj_weight_quantizers.49 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.down_proj_weight_quantizers.50 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.40e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.down_proj_weight_quantizers.51 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.45e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.down_proj_weight_quantizers.52 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.down_proj_weight_quantizers.53 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.down_proj_weight_quantizers.54 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.50e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.down_proj_weight_quantizers.55 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.down_proj_weight_quantizers.56 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.down_proj_weight_quantizers.57 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.down_proj_weight_quantizers.58 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.down_proj_weight_quantizers.59 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.59e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.down_proj_weight_quantizers.60 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.down_proj_weight_quantizers.61 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.down_proj_weight_quantizers.62 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.down_proj_weight_quantizers.63 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.29e-01 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.down_proj_weight_quantizers.64 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.down_proj_weight_quantizers.65 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.down_proj_weight_quantizers.66 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.down_proj_weight_quantizers.67 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.45e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.down_proj_weight_quantizers.68 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.down_proj_weight_quantizers.69 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.40e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.down_proj_weight_quantizers.70 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.down_proj_weight_quantizers.71 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.down_proj_weight_quantizers.72 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.down_proj_weight_quantizers.73 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.45e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.down_proj_weight_quantizers.74 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.down_proj_weight_quantizers.75 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.down_proj_weight_quantizers.76 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.14e-01 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.down_proj_weight_quantizers.77 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.40e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.down_proj_weight_quantizers.78 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.down_proj_weight_quantizers.79 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.45e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.down_proj_weight_quantizers.80 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.down_proj_weight_quantizers.81 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.down_proj_weight_quantizers.82 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.down_proj_weight_quantizers.83 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.down_proj_weight_quantizers.84 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.down_proj_weight_quantizers.85 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.down_proj_weight_quantizers.86 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.down_proj_weight_quantizers.87 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.down_proj_weight_quantizers.88 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.down_proj_weight_quantizers.89 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.50e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.down_proj_weight_quantizers.90 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.45e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.down_proj_weight_quantizers.91 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.down_proj_weight_quantizers.92 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.down_proj_weight_quantizers.93 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.37e-01 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.down_proj_weight_quantizers.94 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.down_proj_weight_quantizers.95 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.down_proj_weight_quantizers.96 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.85e-01 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.down_proj_weight_quantizers.97 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.down_proj_weight_quantizers.98 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.68e-01 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.down_proj_weight_quantizers.99 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.down_proj_weight_quantizers.100 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.down_proj_weight_quantizers.101 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.67e-01 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.down_proj_weight_quantizers.102 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.down_proj_weight_quantizers.103 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.down_proj_weight_quantizers.104 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.down_proj_weight_quantizers.105 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.down_proj_weight_quantizers.106 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.15e-01 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.down_proj_weight_quantizers.107 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.27e-01 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.down_proj_weight_quantizers.108 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.down_proj_weight_quantizers.109 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.down_proj_weight_quantizers.110 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.down_proj_weight_quantizers.111 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.down_proj_weight_quantizers.112 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.15e-01 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.down_proj_weight_quantizers.113 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.54e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.down_proj_weight_quantizers.114 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.down_proj_weight_quantizers.115 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.down_proj_weight_quantizers.116 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.down_proj_weight_quantizers.117 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.down_proj_weight_quantizers.118 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.down_proj_weight_quantizers.119 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.down_proj_weight_quantizers.120 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.down_proj_weight_quantizers.121 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.down_proj_weight_quantizers.122 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.down_proj_weight_quantizers.123 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.down_proj_weight_quantizers.124 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.40e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.down_proj_weight_quantizers.125 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.down_proj_weight_quantizers.126 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.down_proj_weight_quantizers.127 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.down_proj_weight_quantizers.128 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.54e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.down_proj_weight_quantizers.129 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.down_proj_weight_quantizers.130 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.23e-01 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.down_proj_weight_quantizers.131 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.down_proj_weight_quantizers.132 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.64e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.down_proj_weight_quantizers.133 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.41e-01 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.down_proj_weight_quantizers.134 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.down_proj_weight_quantizers.135 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.down_proj_weight_quantizers.136 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.73e-01 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.down_proj_weight_quantizers.137 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.down_proj_weight_quantizers.138 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.down_proj_weight_quantizers.139 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.down_proj_weight_quantizers.140 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.down_proj_weight_quantizers.141 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.down_proj_weight_quantizers.142 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.26e-01 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.down_proj_weight_quantizers.143 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.54e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.down_proj_weight_quantizers.144 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.down_proj_weight_quantizers.145 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.down_proj_weight_quantizers.146 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.down_proj_weight_quantizers.147 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.down_proj_weight_quantizers.148 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.down_proj_weight_quantizers.149 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.down_proj_weight_quantizers.150 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.down_proj_weight_quantizers.151 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.24e-01 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.down_proj_weight_quantizers.152 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.down_proj_weight_quantizers.153 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.down_proj_weight_quantizers.154 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.54e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.down_proj_weight_quantizers.155 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.25e-01 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.down_proj_weight_quantizers.156 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.down_proj_weight_quantizers.157 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.54e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.down_proj_weight_quantizers.158 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.down_proj_weight_quantizers.159 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.down_proj_weight_quantizers.160 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.down_proj_weight_quantizers.161 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.64e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.down_proj_weight_quantizers.162 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.23e-01 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.down_proj_weight_quantizers.163 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.down_proj_weight_quantizers.164 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.down_proj_weight_quantizers.165 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.down_proj_weight_quantizers.166 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.73e-01 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.down_proj_weight_quantizers.167 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.50e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.down_proj_weight_quantizers.168 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.down_proj_weight_quantizers.169 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.down_proj_weight_quantizers.170 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.down_proj_weight_quantizers.171 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.64e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.down_proj_weight_quantizers.172 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.down_proj_weight_quantizers.173 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.down_proj_weight_quantizers.174 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.down_proj_weight_quantizers.175 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.down_proj_weight_quantizers.176 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.54e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.down_proj_weight_quantizers.177 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.down_proj_weight_quantizers.178 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.59e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.down_proj_weight_quantizers.179 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.45e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.down_proj_weight_quantizers.180 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.down_proj_weight_quantizers.181 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.down_proj_weight_quantizers.182 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.down_proj_weight_quantizers.183 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.down_proj_weight_quantizers.184 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.down_proj_weight_quantizers.185 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.down_proj_weight_quantizers.186 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.down_proj_weight_quantizers.187 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.down_proj_weight_quantizers.188 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.down_proj_weight_quantizers.189 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.down_proj_weight_quantizers.190 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.down_proj_weight_quantizers.191 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.down_proj_weight_quantizers.192 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.down_proj_weight_quantizers.193 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.down_proj_weight_quantizers.194 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.26e-01 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.down_proj_weight_quantizers.195 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.down_proj_weight_quantizers.196 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.down_proj_weight_quantizers.197 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.down_proj_weight_quantizers.198 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.down_proj_weight_quantizers.199 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.64e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.down_proj_weight_quantizers.200 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.14e-01 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.down_proj_weight_quantizers.201 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.32e-01 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.down_proj_weight_quantizers.202 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.down_proj_weight_quantizers.203 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.down_proj_weight_quantizers.204 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.down_proj_weight_quantizers.205 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.down_proj_weight_quantizers.206 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.down_proj_weight_quantizers.207 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.15e-01 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.down_proj_weight_quantizers.208 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.down_proj_weight_quantizers.209 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.down_proj_weight_quantizers.210 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.down_proj_weight_quantizers.211 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.down_proj_weight_quantizers.212 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.down_proj_weight_quantizers.213 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.down_proj_weight_quantizers.214 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.down_proj_weight_quantizers.215 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.down_proj_weight_quantizers.216 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.down_proj_weight_quantizers.217 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.43e-01 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.down_proj_weight_quantizers.218 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.down_proj_weight_quantizers.219 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.down_proj_weight_quantizers.220 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.down_proj_weight_quantizers.221 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.down_proj_weight_quantizers.222 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.down_proj_weight_quantizers.223 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.down_proj_weight_quantizers.224 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.40e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.down_proj_weight_quantizers.225 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.down_proj_weight_quantizers.226 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.down_proj_weight_quantizers.227 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.down_proj_weight_quantizers.228 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.down_proj_weight_quantizers.229 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.down_proj_weight_quantizers.230 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.35e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.down_proj_weight_quantizers.231 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.down_proj_weight_quantizers.232 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.down_proj_weight_quantizers.233 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.down_proj_weight_quantizers.234 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.down_proj_weight_quantizers.235 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.down_proj_weight_quantizers.236 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.down_proj_weight_quantizers.237 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.down_proj_weight_quantizers.238 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.down_proj_weight_quantizers.239 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.45e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.down_proj_weight_quantizers.240 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.down_proj_weight_quantizers.241 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.down_proj_weight_quantizers.242 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.down_proj_weight_quantizers.243 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.50e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.down_proj_weight_quantizers.244 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.down_proj_weight_quantizers.245 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.down_proj_weight_quantizers.246 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.32e-01 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.down_proj_weight_quantizers.247 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.38e-01 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.down_proj_weight_quantizers.248 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.down_proj_weight_quantizers.249 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.down_proj_weight_quantizers.250 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.down_proj_weight_quantizers.251 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.down_proj_weight_quantizers.252 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.down_proj_weight_quantizers.253 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.down_proj_weight_quantizers.254 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.down_proj_weight_quantizers.255 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.shared_experts.gate_proj.input_quantizer TensorQuantizer(disabled) +model.layers.26.mlp.shared_experts.gate_proj.output_quantizer TensorQuantizer(disabled) +model.layers.26.mlp.shared_experts.gate_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.26.mlp.shared_experts.up_proj.input_quantizer TensorQuantizer(disabled) +model.layers.26.mlp.shared_experts.up_proj.output_quantizer TensorQuantizer(disabled) +model.layers.26.mlp.shared_experts.up_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.26.mlp.shared_experts.down_proj.input_quantizer TensorQuantizer(disabled) +model.layers.26.mlp.shared_experts.down_proj.output_quantizer TensorQuantizer(disabled) +model.layers.26.mlp.shared_experts.down_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.27.self_attn.q_a_proj.input_quantizer TensorQuantizer(disabled) +model.layers.27.self_attn.q_a_proj.output_quantizer TensorQuantizer(disabled) +model.layers.27.self_attn.q_a_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.27.self_attn.q_b_proj.input_quantizer TensorQuantizer(disabled) +model.layers.27.self_attn.q_b_proj.output_quantizer TensorQuantizer(disabled) +model.layers.27.self_attn.q_b_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.27.self_attn.kv_a_proj_with_mqa.input_quantizer TensorQuantizer(disabled) +model.layers.27.self_attn.kv_a_proj_with_mqa.output_quantizer TensorQuantizer(disabled) +model.layers.27.self_attn.kv_a_proj_with_mqa.weight_quantizer TensorQuantizer(disabled) +model.layers.27.self_attn.kv_b_proj.input_quantizer TensorQuantizer(disabled) +model.layers.27.self_attn.kv_b_proj.output_quantizer TensorQuantizer(disabled) +model.layers.27.self_attn.kv_b_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.27.self_attn.o_proj.input_quantizer TensorQuantizer(disabled) +model.layers.27.self_attn.o_proj.output_quantizer TensorQuantizer(disabled) +model.layers.27.self_attn.o_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.27.self_attn.q_bmm_quantizer TensorQuantizer(disabled) +model.layers.27.self_attn.k_bmm_quantizer TensorQuantizer((4, 3) bit fake per-tensor amax=4.48e+02(const) calibrator=MaxCalibrator quant) +model.layers.27.self_attn.v_bmm_quantizer TensorQuantizer((4, 3) bit fake per-tensor amax=4.48e+02(const) calibrator=MaxCalibrator quant) +model.layers.27.self_attn.p_bmm_quantizer TensorQuantizer(disabled) +model.layers.27.mlp.experts.gate_up_proj_input_quantizer TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e+00 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.gate_up_proj_weight_quantizers.0 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.29e-01 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.gate_up_proj_weight_quantizers.1 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.gate_up_proj_weight_quantizers.2 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.gate_up_proj_weight_quantizers.3 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.gate_up_proj_weight_quantizers.4 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.gate_up_proj_weight_quantizers.5 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.41e-01 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.gate_up_proj_weight_quantizers.6 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.gate_up_proj_weight_quantizers.7 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.gate_up_proj_weight_quantizers.8 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.gate_up_proj_weight_quantizers.9 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.gate_up_proj_weight_quantizers.10 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.gate_up_proj_weight_quantizers.11 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.gate_up_proj_weight_quantizers.12 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.gate_up_proj_weight_quantizers.13 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.gate_up_proj_weight_quantizers.14 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.gate_up_proj_weight_quantizers.15 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.06e-01 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.gate_up_proj_weight_quantizers.16 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.gate_up_proj_weight_quantizers.17 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.gate_up_proj_weight_quantizers.18 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.gate_up_proj_weight_quantizers.19 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.gate_up_proj_weight_quantizers.20 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.gate_up_proj_weight_quantizers.21 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.gate_up_proj_weight_quantizers.22 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.29e-01 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.gate_up_proj_weight_quantizers.23 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.gate_up_proj_weight_quantizers.24 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.gate_up_proj_weight_quantizers.25 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.gate_up_proj_weight_quantizers.26 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.gate_up_proj_weight_quantizers.27 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.gate_up_proj_weight_quantizers.28 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.gate_up_proj_weight_quantizers.29 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.gate_up_proj_weight_quantizers.30 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.32e-01 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.gate_up_proj_weight_quantizers.31 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.gate_up_proj_weight_quantizers.32 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.15e-01 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.gate_up_proj_weight_quantizers.33 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.gate_up_proj_weight_quantizers.34 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.gate_up_proj_weight_quantizers.35 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.gate_up_proj_weight_quantizers.36 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.27e-01 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.gate_up_proj_weight_quantizers.37 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.14e-01 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.gate_up_proj_weight_quantizers.38 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.gate_up_proj_weight_quantizers.39 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.gate_up_proj_weight_quantizers.40 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.gate_up_proj_weight_quantizers.41 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.gate_up_proj_weight_quantizers.42 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.50e-01 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.gate_up_proj_weight_quantizers.43 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.gate_up_proj_weight_quantizers.44 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.gate_up_proj_weight_quantizers.45 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.gate_up_proj_weight_quantizers.46 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.19e-01 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.gate_up_proj_weight_quantizers.47 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.gate_up_proj_weight_quantizers.48 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.gate_up_proj_weight_quantizers.49 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.gate_up_proj_weight_quantizers.50 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.gate_up_proj_weight_quantizers.51 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.gate_up_proj_weight_quantizers.52 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.83e-01 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.gate_up_proj_weight_quantizers.53 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.26e-01 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.gate_up_proj_weight_quantizers.54 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.gate_up_proj_weight_quantizers.55 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.gate_up_proj_weight_quantizers.56 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.gate_up_proj_weight_quantizers.57 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.gate_up_proj_weight_quantizers.58 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.gate_up_proj_weight_quantizers.59 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.15e-01 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.gate_up_proj_weight_quantizers.60 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.34e-01 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.gate_up_proj_weight_quantizers.61 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.gate_up_proj_weight_quantizers.62 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.gate_up_proj_weight_quantizers.63 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.gate_up_proj_weight_quantizers.64 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.gate_up_proj_weight_quantizers.65 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.gate_up_proj_weight_quantizers.66 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.gate_up_proj_weight_quantizers.67 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.17e-01 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.gate_up_proj_weight_quantizers.68 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.gate_up_proj_weight_quantizers.69 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.gate_up_proj_weight_quantizers.70 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.gate_up_proj_weight_quantizers.71 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.33e-01 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.gate_up_proj_weight_quantizers.72 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.13e-01 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.gate_up_proj_weight_quantizers.73 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.22e-01 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.gate_up_proj_weight_quantizers.74 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.gate_up_proj_weight_quantizers.75 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.gate_up_proj_weight_quantizers.76 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.gate_up_proj_weight_quantizers.77 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.gate_up_proj_weight_quantizers.78 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.79e-01 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.gate_up_proj_weight_quantizers.79 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.68e-01 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.gate_up_proj_weight_quantizers.80 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.gate_up_proj_weight_quantizers.81 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.gate_up_proj_weight_quantizers.82 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.17e-01 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.gate_up_proj_weight_quantizers.83 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.gate_up_proj_weight_quantizers.84 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.gate_up_proj_weight_quantizers.85 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.gate_up_proj_weight_quantizers.86 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.gate_up_proj_weight_quantizers.87 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.39e-01 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.gate_up_proj_weight_quantizers.88 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.gate_up_proj_weight_quantizers.89 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.gate_up_proj_weight_quantizers.90 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.gate_up_proj_weight_quantizers.91 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.gate_up_proj_weight_quantizers.92 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.98e-01 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.gate_up_proj_weight_quantizers.93 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.gate_up_proj_weight_quantizers.94 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.gate_up_proj_weight_quantizers.95 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.gate_up_proj_weight_quantizers.96 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.gate_up_proj_weight_quantizers.97 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.gate_up_proj_weight_quantizers.98 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.gate_up_proj_weight_quantizers.99 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.gate_up_proj_weight_quantizers.100 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.gate_up_proj_weight_quantizers.101 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.gate_up_proj_weight_quantizers.102 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.43e-01 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.gate_up_proj_weight_quantizers.103 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.gate_up_proj_weight_quantizers.104 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.gate_up_proj_weight_quantizers.105 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.22e-01 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.gate_up_proj_weight_quantizers.106 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.gate_up_proj_weight_quantizers.107 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.gate_up_proj_weight_quantizers.108 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.gate_up_proj_weight_quantizers.109 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.25e-01 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.gate_up_proj_weight_quantizers.110 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.gate_up_proj_weight_quantizers.111 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.gate_up_proj_weight_quantizers.112 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.gate_up_proj_weight_quantizers.113 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.gate_up_proj_weight_quantizers.114 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.gate_up_proj_weight_quantizers.115 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.51e-01 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.gate_up_proj_weight_quantizers.116 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.gate_up_proj_weight_quantizers.117 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.gate_up_proj_weight_quantizers.118 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.gate_up_proj_weight_quantizers.119 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.gate_up_proj_weight_quantizers.120 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.gate_up_proj_weight_quantizers.121 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.gate_up_proj_weight_quantizers.122 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.gate_up_proj_weight_quantizers.123 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.gate_up_proj_weight_quantizers.124 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.gate_up_proj_weight_quantizers.125 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.02e-01 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.gate_up_proj_weight_quantizers.126 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.gate_up_proj_weight_quantizers.127 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.gate_up_proj_weight_quantizers.128 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.gate_up_proj_weight_quantizers.129 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.gate_up_proj_weight_quantizers.130 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.gate_up_proj_weight_quantizers.131 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.gate_up_proj_weight_quantizers.132 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.gate_up_proj_weight_quantizers.133 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.gate_up_proj_weight_quantizers.134 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.gate_up_proj_weight_quantizers.135 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.gate_up_proj_weight_quantizers.136 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.gate_up_proj_weight_quantizers.137 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.17e-01 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.gate_up_proj_weight_quantizers.138 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.gate_up_proj_weight_quantizers.139 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.gate_up_proj_weight_quantizers.140 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.gate_up_proj_weight_quantizers.141 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.28e-01 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.gate_up_proj_weight_quantizers.142 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.gate_up_proj_weight_quantizers.143 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.gate_up_proj_weight_quantizers.144 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.gate_up_proj_weight_quantizers.145 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.16e-01 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.gate_up_proj_weight_quantizers.146 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.gate_up_proj_weight_quantizers.147 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.gate_up_proj_weight_quantizers.148 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.gate_up_proj_weight_quantizers.149 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.gate_up_proj_weight_quantizers.150 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.gate_up_proj_weight_quantizers.151 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.03e-01 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.gate_up_proj_weight_quantizers.152 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.gate_up_proj_weight_quantizers.153 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.gate_up_proj_weight_quantizers.154 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.30e-01 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.gate_up_proj_weight_quantizers.155 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.83e-01 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.gate_up_proj_weight_quantizers.156 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.gate_up_proj_weight_quantizers.157 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.gate_up_proj_weight_quantizers.158 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.gate_up_proj_weight_quantizers.159 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.gate_up_proj_weight_quantizers.160 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.26e-01 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.gate_up_proj_weight_quantizers.161 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.gate_up_proj_weight_quantizers.162 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.gate_up_proj_weight_quantizers.163 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.14e-01 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.gate_up_proj_weight_quantizers.164 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.gate_up_proj_weight_quantizers.165 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.20e-01 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.gate_up_proj_weight_quantizers.166 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.gate_up_proj_weight_quantizers.167 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.gate_up_proj_weight_quantizers.168 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.86e-01 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.gate_up_proj_weight_quantizers.169 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.gate_up_proj_weight_quantizers.170 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.67e-01 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.gate_up_proj_weight_quantizers.171 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.13e-01 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.gate_up_proj_weight_quantizers.172 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.gate_up_proj_weight_quantizers.173 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.38e-01 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.gate_up_proj_weight_quantizers.174 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.gate_up_proj_weight_quantizers.175 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.gate_up_proj_weight_quantizers.176 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.gate_up_proj_weight_quantizers.177 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.gate_up_proj_weight_quantizers.178 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.28e-01 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.gate_up_proj_weight_quantizers.179 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.gate_up_proj_weight_quantizers.180 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.gate_up_proj_weight_quantizers.181 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.gate_up_proj_weight_quantizers.182 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.gate_up_proj_weight_quantizers.183 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.gate_up_proj_weight_quantizers.184 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.gate_up_proj_weight_quantizers.185 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.gate_up_proj_weight_quantizers.186 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.gate_up_proj_weight_quantizers.187 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.gate_up_proj_weight_quantizers.188 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.52e-01 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.gate_up_proj_weight_quantizers.189 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.gate_up_proj_weight_quantizers.190 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.gate_up_proj_weight_quantizers.191 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.gate_up_proj_weight_quantizers.192 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.gate_up_proj_weight_quantizers.193 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.gate_up_proj_weight_quantizers.194 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.gate_up_proj_weight_quantizers.195 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.gate_up_proj_weight_quantizers.196 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.gate_up_proj_weight_quantizers.197 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.gate_up_proj_weight_quantizers.198 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.gate_up_proj_weight_quantizers.199 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.gate_up_proj_weight_quantizers.200 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.gate_up_proj_weight_quantizers.201 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.79e-01 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.gate_up_proj_weight_quantizers.202 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.21e-01 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.gate_up_proj_weight_quantizers.203 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.gate_up_proj_weight_quantizers.204 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.gate_up_proj_weight_quantizers.205 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.gate_up_proj_weight_quantizers.206 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.gate_up_proj_weight_quantizers.207 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.gate_up_proj_weight_quantizers.208 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.gate_up_proj_weight_quantizers.209 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.gate_up_proj_weight_quantizers.210 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.gate_up_proj_weight_quantizers.211 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.73e-01 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.gate_up_proj_weight_quantizers.212 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.gate_up_proj_weight_quantizers.213 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.gate_up_proj_weight_quantizers.214 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.gate_up_proj_weight_quantizers.215 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.gate_up_proj_weight_quantizers.216 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.15e-01 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.gate_up_proj_weight_quantizers.217 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.gate_up_proj_weight_quantizers.218 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.35e-01 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.gate_up_proj_weight_quantizers.219 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.18e-01 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.gate_up_proj_weight_quantizers.220 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.gate_up_proj_weight_quantizers.221 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.gate_up_proj_weight_quantizers.222 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.gate_up_proj_weight_quantizers.223 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.gate_up_proj_weight_quantizers.224 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.15e-01 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.gate_up_proj_weight_quantizers.225 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.20e-01 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.gate_up_proj_weight_quantizers.226 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.gate_up_proj_weight_quantizers.227 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.gate_up_proj_weight_quantizers.228 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.gate_up_proj_weight_quantizers.229 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.gate_up_proj_weight_quantizers.230 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.gate_up_proj_weight_quantizers.231 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.28e-01 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.gate_up_proj_weight_quantizers.232 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.gate_up_proj_weight_quantizers.233 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.23e-01 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.gate_up_proj_weight_quantizers.234 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.gate_up_proj_weight_quantizers.235 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.14e-01 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.gate_up_proj_weight_quantizers.236 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.gate_up_proj_weight_quantizers.237 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.gate_up_proj_weight_quantizers.238 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.gate_up_proj_weight_quantizers.239 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.gate_up_proj_weight_quantizers.240 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.gate_up_proj_weight_quantizers.241 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.19e-01 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.gate_up_proj_weight_quantizers.242 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.gate_up_proj_weight_quantizers.243 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.gate_up_proj_weight_quantizers.244 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.gate_up_proj_weight_quantizers.245 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.gate_up_proj_weight_quantizers.246 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.gate_up_proj_weight_quantizers.247 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.30e-01 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.gate_up_proj_weight_quantizers.248 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.51e-01 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.gate_up_proj_weight_quantizers.249 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.27e-01 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.gate_up_proj_weight_quantizers.250 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.gate_up_proj_weight_quantizers.251 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.gate_up_proj_weight_quantizers.252 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.gate_up_proj_weight_quantizers.253 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.gate_up_proj_weight_quantizers.254 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.gate_up_proj_weight_quantizers.255 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.76e-01 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.down_proj_input_quantizer TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.12e+00 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.down_proj_weight_quantizers.0 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.19e-01 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.down_proj_weight_quantizers.1 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.down_proj_weight_quantizers.2 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.down_proj_weight_quantizers.3 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.down_proj_weight_quantizers.4 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.down_proj_weight_quantizers.5 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.down_proj_weight_quantizers.6 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.down_proj_weight_quantizers.7 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.down_proj_weight_quantizers.8 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.down_proj_weight_quantizers.9 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.down_proj_weight_quantizers.10 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.20e-01 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.down_proj_weight_quantizers.11 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.down_proj_weight_quantizers.12 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.down_proj_weight_quantizers.13 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.down_proj_weight_quantizers.14 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.down_proj_weight_quantizers.15 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.down_proj_weight_quantizers.16 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.down_proj_weight_quantizers.17 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.82e-01 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.down_proj_weight_quantizers.18 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.20e-01 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.down_proj_weight_quantizers.19 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.down_proj_weight_quantizers.20 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.down_proj_weight_quantizers.21 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.down_proj_weight_quantizers.22 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.45e-01 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.down_proj_weight_quantizers.23 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.down_proj_weight_quantizers.24 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.down_proj_weight_quantizers.25 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.down_proj_weight_quantizers.26 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.down_proj_weight_quantizers.27 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.down_proj_weight_quantizers.28 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.down_proj_weight_quantizers.29 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.down_proj_weight_quantizers.30 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.down_proj_weight_quantizers.31 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.down_proj_weight_quantizers.32 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.37e-01 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.down_proj_weight_quantizers.33 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.down_proj_weight_quantizers.34 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.down_proj_weight_quantizers.35 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.down_proj_weight_quantizers.36 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.down_proj_weight_quantizers.37 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.down_proj_weight_quantizers.38 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.down_proj_weight_quantizers.39 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.down_proj_weight_quantizers.40 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.down_proj_weight_quantizers.41 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.down_proj_weight_quantizers.42 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.down_proj_weight_quantizers.43 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.down_proj_weight_quantizers.44 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.down_proj_weight_quantizers.45 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.down_proj_weight_quantizers.46 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.down_proj_weight_quantizers.47 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.40e-02 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.down_proj_weight_quantizers.48 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.50e-02 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.down_proj_weight_quantizers.49 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.down_proj_weight_quantizers.50 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.down_proj_weight_quantizers.51 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.down_proj_weight_quantizers.52 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.62e-01 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.down_proj_weight_quantizers.53 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.64e-02 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.down_proj_weight_quantizers.54 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.down_proj_weight_quantizers.55 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.down_proj_weight_quantizers.56 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.down_proj_weight_quantizers.57 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.down_proj_weight_quantizers.58 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.down_proj_weight_quantizers.59 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.down_proj_weight_quantizers.60 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.30e-01 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.down_proj_weight_quantizers.61 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.down_proj_weight_quantizers.62 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.down_proj_weight_quantizers.63 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.36e-01 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.down_proj_weight_quantizers.64 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.down_proj_weight_quantizers.65 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.down_proj_weight_quantizers.66 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.down_proj_weight_quantizers.67 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.down_proj_weight_quantizers.68 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.down_proj_weight_quantizers.69 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.35e-02 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.down_proj_weight_quantizers.70 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.down_proj_weight_quantizers.71 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.down_proj_weight_quantizers.72 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.down_proj_weight_quantizers.73 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.30e-01 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.down_proj_weight_quantizers.74 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.down_proj_weight_quantizers.75 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.59e-02 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.down_proj_weight_quantizers.76 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.20e-01 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.down_proj_weight_quantizers.77 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.down_proj_weight_quantizers.78 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.22e-01 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.down_proj_weight_quantizers.79 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.64e-01 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.down_proj_weight_quantizers.80 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.down_proj_weight_quantizers.81 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.down_proj_weight_quantizers.82 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.down_proj_weight_quantizers.83 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.down_proj_weight_quantizers.84 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.down_proj_weight_quantizers.85 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.64e-02 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.down_proj_weight_quantizers.86 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.down_proj_weight_quantizers.87 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.down_proj_weight_quantizers.88 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.13e-01 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.down_proj_weight_quantizers.89 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.down_proj_weight_quantizers.90 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.down_proj_weight_quantizers.91 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.down_proj_weight_quantizers.92 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.down_proj_weight_quantizers.93 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.22e-01 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.down_proj_weight_quantizers.94 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.down_proj_weight_quantizers.95 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.down_proj_weight_quantizers.96 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.down_proj_weight_quantizers.97 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.down_proj_weight_quantizers.98 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.down_proj_weight_quantizers.99 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.28e-01 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.down_proj_weight_quantizers.100 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.down_proj_weight_quantizers.101 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.down_proj_weight_quantizers.102 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.down_proj_weight_quantizers.103 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.down_proj_weight_quantizers.104 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.down_proj_weight_quantizers.105 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.89e-01 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.down_proj_weight_quantizers.106 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.down_proj_weight_quantizers.107 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.down_proj_weight_quantizers.108 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.down_proj_weight_quantizers.109 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.23e-01 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.down_proj_weight_quantizers.110 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.down_proj_weight_quantizers.111 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.down_proj_weight_quantizers.112 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.down_proj_weight_quantizers.113 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.48e-01 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.down_proj_weight_quantizers.114 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.down_proj_weight_quantizers.115 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.62e-01 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.down_proj_weight_quantizers.116 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.down_proj_weight_quantizers.117 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.down_proj_weight_quantizers.118 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.19e-01 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.down_proj_weight_quantizers.119 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.down_proj_weight_quantizers.120 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.54e-02 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.down_proj_weight_quantizers.121 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.down_proj_weight_quantizers.122 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.down_proj_weight_quantizers.123 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.38e-01 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.down_proj_weight_quantizers.124 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.down_proj_weight_quantizers.125 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.44e-01 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.down_proj_weight_quantizers.126 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.down_proj_weight_quantizers.127 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.down_proj_weight_quantizers.128 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.down_proj_weight_quantizers.129 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.down_proj_weight_quantizers.130 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.down_proj_weight_quantizers.131 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.down_proj_weight_quantizers.132 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.down_proj_weight_quantizers.133 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.54e-02 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.down_proj_weight_quantizers.134 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.down_proj_weight_quantizers.135 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.down_proj_weight_quantizers.136 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.down_proj_weight_quantizers.137 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.21e-01 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.down_proj_weight_quantizers.138 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.down_proj_weight_quantizers.139 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.21e-01 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.down_proj_weight_quantizers.140 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.down_proj_weight_quantizers.141 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.38e-01 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.down_proj_weight_quantizers.142 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.down_proj_weight_quantizers.143 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.17e-01 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.down_proj_weight_quantizers.144 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.down_proj_weight_quantizers.145 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.down_proj_weight_quantizers.146 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.down_proj_weight_quantizers.147 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.down_proj_weight_quantizers.148 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.57e-01 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.down_proj_weight_quantizers.149 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.down_proj_weight_quantizers.150 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.down_proj_weight_quantizers.151 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.99e-01 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.down_proj_weight_quantizers.152 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.down_proj_weight_quantizers.153 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.45e-02 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.down_proj_weight_quantizers.154 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.down_proj_weight_quantizers.155 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.down_proj_weight_quantizers.156 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.30e-01 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.down_proj_weight_quantizers.157 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.40e-02 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.down_proj_weight_quantizers.158 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.down_proj_weight_quantizers.159 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.down_proj_weight_quantizers.160 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.down_proj_weight_quantizers.161 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.down_proj_weight_quantizers.162 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.down_proj_weight_quantizers.163 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.down_proj_weight_quantizers.164 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.down_proj_weight_quantizers.165 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.down_proj_weight_quantizers.166 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.down_proj_weight_quantizers.167 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.down_proj_weight_quantizers.168 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.99e-01 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.down_proj_weight_quantizers.169 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.59e-02 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.down_proj_weight_quantizers.170 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.96e-01 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.down_proj_weight_quantizers.171 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.down_proj_weight_quantizers.172 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.61e-01 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.down_proj_weight_quantizers.173 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.down_proj_weight_quantizers.174 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.down_proj_weight_quantizers.175 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.59e-02 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.down_proj_weight_quantizers.176 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.down_proj_weight_quantizers.177 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.down_proj_weight_quantizers.178 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.22e-01 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.down_proj_weight_quantizers.179 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.down_proj_weight_quantizers.180 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.down_proj_weight_quantizers.181 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.down_proj_weight_quantizers.182 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.59e-02 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.down_proj_weight_quantizers.183 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.down_proj_weight_quantizers.184 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.down_proj_weight_quantizers.185 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.down_proj_weight_quantizers.186 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.down_proj_weight_quantizers.187 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.down_proj_weight_quantizers.188 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.25e-01 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.down_proj_weight_quantizers.189 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.50e-02 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.down_proj_weight_quantizers.190 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.down_proj_weight_quantizers.191 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.down_proj_weight_quantizers.192 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.down_proj_weight_quantizers.193 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.down_proj_weight_quantizers.194 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.down_proj_weight_quantizers.195 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.down_proj_weight_quantizers.196 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.down_proj_weight_quantizers.197 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.down_proj_weight_quantizers.198 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.down_proj_weight_quantizers.199 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.down_proj_weight_quantizers.200 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.64e-02 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.down_proj_weight_quantizers.201 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.97e-01 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.down_proj_weight_quantizers.202 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.68e-01 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.down_proj_weight_quantizers.203 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.49e-01 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.down_proj_weight_quantizers.204 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.down_proj_weight_quantizers.205 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.down_proj_weight_quantizers.206 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.down_proj_weight_quantizers.207 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.38e-01 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.down_proj_weight_quantizers.208 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.down_proj_weight_quantizers.209 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.93e-01 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.down_proj_weight_quantizers.210 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.49e-01 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.down_proj_weight_quantizers.211 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.24e-01 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.down_proj_weight_quantizers.212 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.down_proj_weight_quantizers.213 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.down_proj_weight_quantizers.214 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.down_proj_weight_quantizers.215 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.down_proj_weight_quantizers.216 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.down_proj_weight_quantizers.217 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.16e-01 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.down_proj_weight_quantizers.218 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.down_proj_weight_quantizers.219 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.97e-01 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.down_proj_weight_quantizers.220 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.down_proj_weight_quantizers.221 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.down_proj_weight_quantizers.222 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.down_proj_weight_quantizers.223 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.15e-01 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.down_proj_weight_quantizers.224 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.17e-01 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.down_proj_weight_quantizers.225 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.down_proj_weight_quantizers.226 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.down_proj_weight_quantizers.227 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.down_proj_weight_quantizers.228 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.35e-02 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.down_proj_weight_quantizers.229 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.down_proj_weight_quantizers.230 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.down_proj_weight_quantizers.231 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.down_proj_weight_quantizers.232 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.down_proj_weight_quantizers.233 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.down_proj_weight_quantizers.234 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.50e-02 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.down_proj_weight_quantizers.235 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.46e-01 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.down_proj_weight_quantizers.236 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.down_proj_weight_quantizers.237 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.down_proj_weight_quantizers.238 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.down_proj_weight_quantizers.239 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.down_proj_weight_quantizers.240 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.down_proj_weight_quantizers.241 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.22e-01 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.down_proj_weight_quantizers.242 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.59e-02 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.down_proj_weight_quantizers.243 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.30e-02 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.down_proj_weight_quantizers.244 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.59e-02 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.down_proj_weight_quantizers.245 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.45e-02 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.down_proj_weight_quantizers.246 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.down_proj_weight_quantizers.247 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.22e-01 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.down_proj_weight_quantizers.248 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.down_proj_weight_quantizers.249 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.48e-01 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.down_proj_weight_quantizers.250 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.down_proj_weight_quantizers.251 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.down_proj_weight_quantizers.252 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.down_proj_weight_quantizers.253 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.down_proj_weight_quantizers.254 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.64e-02 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.down_proj_weight_quantizers.255 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.13e-01 calibrator=MaxCalibrator quant) +model.layers.27.mlp.shared_experts.gate_proj.input_quantizer TensorQuantizer(disabled) +model.layers.27.mlp.shared_experts.gate_proj.output_quantizer TensorQuantizer(disabled) +model.layers.27.mlp.shared_experts.gate_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.27.mlp.shared_experts.up_proj.input_quantizer TensorQuantizer(disabled) +model.layers.27.mlp.shared_experts.up_proj.output_quantizer TensorQuantizer(disabled) +model.layers.27.mlp.shared_experts.up_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.27.mlp.shared_experts.down_proj.input_quantizer TensorQuantizer(disabled) +model.layers.27.mlp.shared_experts.down_proj.output_quantizer TensorQuantizer(disabled) +model.layers.27.mlp.shared_experts.down_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.28.self_attn.q_a_proj.input_quantizer TensorQuantizer(disabled) +model.layers.28.self_attn.q_a_proj.output_quantizer TensorQuantizer(disabled) +model.layers.28.self_attn.q_a_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.28.self_attn.q_b_proj.input_quantizer TensorQuantizer(disabled) +model.layers.28.self_attn.q_b_proj.output_quantizer TensorQuantizer(disabled) +model.layers.28.self_attn.q_b_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.28.self_attn.kv_a_proj_with_mqa.input_quantizer TensorQuantizer(disabled) +model.layers.28.self_attn.kv_a_proj_with_mqa.output_quantizer TensorQuantizer(disabled) +model.layers.28.self_attn.kv_a_proj_with_mqa.weight_quantizer TensorQuantizer(disabled) +model.layers.28.self_attn.kv_b_proj.input_quantizer TensorQuantizer(disabled) +model.layers.28.self_attn.kv_b_proj.output_quantizer TensorQuantizer(disabled) +model.layers.28.self_attn.kv_b_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.28.self_attn.o_proj.input_quantizer TensorQuantizer(disabled) +model.layers.28.self_attn.o_proj.output_quantizer TensorQuantizer(disabled) +model.layers.28.self_attn.o_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.28.self_attn.q_bmm_quantizer TensorQuantizer(disabled) +model.layers.28.self_attn.k_bmm_quantizer TensorQuantizer((4, 3) bit fake per-tensor amax=4.48e+02(const) calibrator=MaxCalibrator quant) +model.layers.28.self_attn.v_bmm_quantizer TensorQuantizer((4, 3) bit fake per-tensor amax=4.48e+02(const) calibrator=MaxCalibrator quant) +model.layers.28.self_attn.p_bmm_quantizer TensorQuantizer(disabled) +model.layers.28.mlp.experts.gate_up_proj_input_quantizer TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e+00 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.gate_up_proj_weight_quantizers.0 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.gate_up_proj_weight_quantizers.1 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.gate_up_proj_weight_quantizers.2 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.65e-01 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.gate_up_proj_weight_quantizers.3 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.gate_up_proj_weight_quantizers.4 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.gate_up_proj_weight_quantizers.5 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.31e-01 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.gate_up_proj_weight_quantizers.6 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.gate_up_proj_weight_quantizers.7 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.21e-01 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.gate_up_proj_weight_quantizers.8 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.15e-01 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.gate_up_proj_weight_quantizers.9 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.gate_up_proj_weight_quantizers.10 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.gate_up_proj_weight_quantizers.11 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.46e-01 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.gate_up_proj_weight_quantizers.12 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.gate_up_proj_weight_quantizers.13 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.gate_up_proj_weight_quantizers.14 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.gate_up_proj_weight_quantizers.15 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.gate_up_proj_weight_quantizers.16 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.gate_up_proj_weight_quantizers.17 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.gate_up_proj_weight_quantizers.18 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.gate_up_proj_weight_quantizers.19 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.gate_up_proj_weight_quantizers.20 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.42e-01 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.gate_up_proj_weight_quantizers.21 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.36e-01 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.gate_up_proj_weight_quantizers.22 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.gate_up_proj_weight_quantizers.23 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.gate_up_proj_weight_quantizers.24 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.gate_up_proj_weight_quantizers.25 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.gate_up_proj_weight_quantizers.26 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.26e-01 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.gate_up_proj_weight_quantizers.27 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.gate_up_proj_weight_quantizers.28 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.04e-01 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.gate_up_proj_weight_quantizers.29 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.gate_up_proj_weight_quantizers.30 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.gate_up_proj_weight_quantizers.31 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.gate_up_proj_weight_quantizers.32 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.gate_up_proj_weight_quantizers.33 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.14e-01 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.gate_up_proj_weight_quantizers.34 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.gate_up_proj_weight_quantizers.35 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.gate_up_proj_weight_quantizers.36 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.91e-01 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.gate_up_proj_weight_quantizers.37 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.gate_up_proj_weight_quantizers.38 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.gate_up_proj_weight_quantizers.39 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.gate_up_proj_weight_quantizers.40 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.gate_up_proj_weight_quantizers.41 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.gate_up_proj_weight_quantizers.42 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.gate_up_proj_weight_quantizers.43 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.gate_up_proj_weight_quantizers.44 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.gate_up_proj_weight_quantizers.45 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.66e-01 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.gate_up_proj_weight_quantizers.46 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.gate_up_proj_weight_quantizers.47 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.17e-01 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.gate_up_proj_weight_quantizers.48 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.gate_up_proj_weight_quantizers.49 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.gate_up_proj_weight_quantizers.50 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.19e-01 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.gate_up_proj_weight_quantizers.51 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.gate_up_proj_weight_quantizers.52 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.gate_up_proj_weight_quantizers.53 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.gate_up_proj_weight_quantizers.54 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.gate_up_proj_weight_quantizers.55 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.gate_up_proj_weight_quantizers.56 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.gate_up_proj_weight_quantizers.57 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.gate_up_proj_weight_quantizers.58 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.gate_up_proj_weight_quantizers.59 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.32e-01 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.gate_up_proj_weight_quantizers.60 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.gate_up_proj_weight_quantizers.61 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.gate_up_proj_weight_quantizers.62 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.gate_up_proj_weight_quantizers.63 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.gate_up_proj_weight_quantizers.64 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.gate_up_proj_weight_quantizers.65 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.gate_up_proj_weight_quantizers.66 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.gate_up_proj_weight_quantizers.67 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.gate_up_proj_weight_quantizers.68 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.gate_up_proj_weight_quantizers.69 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.gate_up_proj_weight_quantizers.70 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.gate_up_proj_weight_quantizers.71 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.gate_up_proj_weight_quantizers.72 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.gate_up_proj_weight_quantizers.73 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.gate_up_proj_weight_quantizers.74 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.gate_up_proj_weight_quantizers.75 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.gate_up_proj_weight_quantizers.76 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.gate_up_proj_weight_quantizers.77 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.gate_up_proj_weight_quantizers.78 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.gate_up_proj_weight_quantizers.79 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.34e-01 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.gate_up_proj_weight_quantizers.80 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.gate_up_proj_weight_quantizers.81 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.gate_up_proj_weight_quantizers.82 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.32e-01 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.gate_up_proj_weight_quantizers.83 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.35e-01 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.gate_up_proj_weight_quantizers.84 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.gate_up_proj_weight_quantizers.85 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.gate_up_proj_weight_quantizers.86 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.gate_up_proj_weight_quantizers.87 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.27e-01 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.gate_up_proj_weight_quantizers.88 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.gate_up_proj_weight_quantizers.89 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.gate_up_proj_weight_quantizers.90 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.gate_up_proj_weight_quantizers.91 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.gate_up_proj_weight_quantizers.92 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.gate_up_proj_weight_quantizers.93 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.gate_up_proj_weight_quantizers.94 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.gate_up_proj_weight_quantizers.95 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.gate_up_proj_weight_quantizers.96 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.gate_up_proj_weight_quantizers.97 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.gate_up_proj_weight_quantizers.98 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.gate_up_proj_weight_quantizers.99 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.gate_up_proj_weight_quantizers.100 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.gate_up_proj_weight_quantizers.101 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.gate_up_proj_weight_quantizers.102 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.gate_up_proj_weight_quantizers.103 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.gate_up_proj_weight_quantizers.104 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.58e-01 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.gate_up_proj_weight_quantizers.105 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.gate_up_proj_weight_quantizers.106 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.gate_up_proj_weight_quantizers.107 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.gate_up_proj_weight_quantizers.108 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.gate_up_proj_weight_quantizers.109 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.gate_up_proj_weight_quantizers.110 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.gate_up_proj_weight_quantizers.111 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.gate_up_proj_weight_quantizers.112 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.gate_up_proj_weight_quantizers.113 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.gate_up_proj_weight_quantizers.114 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.68e-01 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.gate_up_proj_weight_quantizers.115 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.05e-01 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.gate_up_proj_weight_quantizers.116 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.gate_up_proj_weight_quantizers.117 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.gate_up_proj_weight_quantizers.118 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.gate_up_proj_weight_quantizers.119 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.gate_up_proj_weight_quantizers.120 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.gate_up_proj_weight_quantizers.121 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.gate_up_proj_weight_quantizers.122 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.gate_up_proj_weight_quantizers.123 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.19e-01 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.gate_up_proj_weight_quantizers.124 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.gate_up_proj_weight_quantizers.125 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.gate_up_proj_weight_quantizers.126 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.gate_up_proj_weight_quantizers.127 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.gate_up_proj_weight_quantizers.128 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.gate_up_proj_weight_quantizers.129 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.gate_up_proj_weight_quantizers.130 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.37e-01 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.gate_up_proj_weight_quantizers.131 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.gate_up_proj_weight_quantizers.132 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.gate_up_proj_weight_quantizers.133 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.gate_up_proj_weight_quantizers.134 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.gate_up_proj_weight_quantizers.135 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.gate_up_proj_weight_quantizers.136 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.gate_up_proj_weight_quantizers.137 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.gate_up_proj_weight_quantizers.138 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.gate_up_proj_weight_quantizers.139 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.gate_up_proj_weight_quantizers.140 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.gate_up_proj_weight_quantizers.141 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.gate_up_proj_weight_quantizers.142 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.gate_up_proj_weight_quantizers.143 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.gate_up_proj_weight_quantizers.144 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.gate_up_proj_weight_quantizers.145 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.gate_up_proj_weight_quantizers.146 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.gate_up_proj_weight_quantizers.147 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.24e-01 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.gate_up_proj_weight_quantizers.148 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.gate_up_proj_weight_quantizers.149 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.39e-01 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.gate_up_proj_weight_quantizers.150 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.gate_up_proj_weight_quantizers.151 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.gate_up_proj_weight_quantizers.152 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.21e-01 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.gate_up_proj_weight_quantizers.153 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.gate_up_proj_weight_quantizers.154 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.gate_up_proj_weight_quantizers.155 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.gate_up_proj_weight_quantizers.156 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.gate_up_proj_weight_quantizers.157 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.32e-01 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.gate_up_proj_weight_quantizers.158 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.gate_up_proj_weight_quantizers.159 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.gate_up_proj_weight_quantizers.160 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.gate_up_proj_weight_quantizers.161 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.gate_up_proj_weight_quantizers.162 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.gate_up_proj_weight_quantizers.163 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.06e-01 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.gate_up_proj_weight_quantizers.164 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.65e-01 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.gate_up_proj_weight_quantizers.165 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.gate_up_proj_weight_quantizers.166 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.15e-01 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.gate_up_proj_weight_quantizers.167 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.gate_up_proj_weight_quantizers.168 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.gate_up_proj_weight_quantizers.169 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.56e-01 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.gate_up_proj_weight_quantizers.170 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.gate_up_proj_weight_quantizers.171 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.48e-01 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.gate_up_proj_weight_quantizers.172 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.gate_up_proj_weight_quantizers.173 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.87e-01 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.gate_up_proj_weight_quantizers.174 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.gate_up_proj_weight_quantizers.175 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.34e-01 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.gate_up_proj_weight_quantizers.176 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.gate_up_proj_weight_quantizers.177 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.gate_up_proj_weight_quantizers.178 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.gate_up_proj_weight_quantizers.179 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.gate_up_proj_weight_quantizers.180 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.gate_up_proj_weight_quantizers.181 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.gate_up_proj_weight_quantizers.182 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.gate_up_proj_weight_quantizers.183 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.gate_up_proj_weight_quantizers.184 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.18e-01 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.gate_up_proj_weight_quantizers.185 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.15e-01 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.gate_up_proj_weight_quantizers.186 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.35e-01 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.gate_up_proj_weight_quantizers.187 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.gate_up_proj_weight_quantizers.188 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.41e-01 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.gate_up_proj_weight_quantizers.189 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.gate_up_proj_weight_quantizers.190 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.gate_up_proj_weight_quantizers.191 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.gate_up_proj_weight_quantizers.192 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.64e-02 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.gate_up_proj_weight_quantizers.193 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.gate_up_proj_weight_quantizers.194 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.gate_up_proj_weight_quantizers.195 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.gate_up_proj_weight_quantizers.196 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.gate_up_proj_weight_quantizers.197 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.gate_up_proj_weight_quantizers.198 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.gate_up_proj_weight_quantizers.199 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.93e-01 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.gate_up_proj_weight_quantizers.200 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.gate_up_proj_weight_quantizers.201 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.gate_up_proj_weight_quantizers.202 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.gate_up_proj_weight_quantizers.203 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.gate_up_proj_weight_quantizers.204 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.56e-01 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.gate_up_proj_weight_quantizers.205 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.gate_up_proj_weight_quantizers.206 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.gate_up_proj_weight_quantizers.207 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.gate_up_proj_weight_quantizers.208 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.gate_up_proj_weight_quantizers.209 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.gate_up_proj_weight_quantizers.210 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.gate_up_proj_weight_quantizers.211 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.16e-01 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.gate_up_proj_weight_quantizers.212 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.gate_up_proj_weight_quantizers.213 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.gate_up_proj_weight_quantizers.214 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.gate_up_proj_weight_quantizers.215 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.23e-01 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.gate_up_proj_weight_quantizers.216 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.gate_up_proj_weight_quantizers.217 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.41e-01 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.gate_up_proj_weight_quantizers.218 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.gate_up_proj_weight_quantizers.219 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.gate_up_proj_weight_quantizers.220 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.gate_up_proj_weight_quantizers.221 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.gate_up_proj_weight_quantizers.222 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.gate_up_proj_weight_quantizers.223 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.gate_up_proj_weight_quantizers.224 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.83e-01 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.gate_up_proj_weight_quantizers.225 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.gate_up_proj_weight_quantizers.226 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.gate_up_proj_weight_quantizers.227 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.gate_up_proj_weight_quantizers.228 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.gate_up_proj_weight_quantizers.229 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.79e-01 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.gate_up_proj_weight_quantizers.230 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.gate_up_proj_weight_quantizers.231 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.gate_up_proj_weight_quantizers.232 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.gate_up_proj_weight_quantizers.233 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.gate_up_proj_weight_quantizers.234 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.gate_up_proj_weight_quantizers.235 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.gate_up_proj_weight_quantizers.236 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.gate_up_proj_weight_quantizers.237 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.gate_up_proj_weight_quantizers.238 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.gate_up_proj_weight_quantizers.239 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.15e-01 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.gate_up_proj_weight_quantizers.240 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.gate_up_proj_weight_quantizers.241 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.gate_up_proj_weight_quantizers.242 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.44e-01 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.gate_up_proj_weight_quantizers.243 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.gate_up_proj_weight_quantizers.244 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.57e-01 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.gate_up_proj_weight_quantizers.245 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.gate_up_proj_weight_quantizers.246 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.gate_up_proj_weight_quantizers.247 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.gate_up_proj_weight_quantizers.248 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.23e-01 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.gate_up_proj_weight_quantizers.249 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.gate_up_proj_weight_quantizers.250 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.gate_up_proj_weight_quantizers.251 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.gate_up_proj_weight_quantizers.252 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.gate_up_proj_weight_quantizers.253 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.gate_up_proj_weight_quantizers.254 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.gate_up_proj_weight_quantizers.255 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.13e-01 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.down_proj_input_quantizer TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.02e+01 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.down_proj_weight_quantizers.0 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.30e-02 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.down_proj_weight_quantizers.1 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.down_proj_weight_quantizers.2 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.35e-01 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.down_proj_weight_quantizers.3 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.14e-01 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.down_proj_weight_quantizers.4 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.59e-02 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.down_proj_weight_quantizers.5 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.23e-01 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.down_proj_weight_quantizers.6 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.down_proj_weight_quantizers.7 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.18e-01 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.down_proj_weight_quantizers.8 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.27e-01 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.down_proj_weight_quantizers.9 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.59e-02 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.down_proj_weight_quantizers.10 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.down_proj_weight_quantizers.11 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.18e-01 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.down_proj_weight_quantizers.12 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.down_proj_weight_quantizers.13 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.down_proj_weight_quantizers.14 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.50e-02 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.down_proj_weight_quantizers.15 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.down_proj_weight_quantizers.16 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.down_proj_weight_quantizers.17 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.down_proj_weight_quantizers.18 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.down_proj_weight_quantizers.19 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.down_proj_weight_quantizers.20 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.down_proj_weight_quantizers.21 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.down_proj_weight_quantizers.22 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.down_proj_weight_quantizers.23 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.59e-02 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.down_proj_weight_quantizers.24 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.down_proj_weight_quantizers.25 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.down_proj_weight_quantizers.26 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.down_proj_weight_quantizers.27 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.50e-02 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.down_proj_weight_quantizers.28 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.46e-01 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.down_proj_weight_quantizers.29 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.down_proj_weight_quantizers.30 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.54e-02 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.down_proj_weight_quantizers.31 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.20e-02 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.down_proj_weight_quantizers.32 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.down_proj_weight_quantizers.33 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.35e-02 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.down_proj_weight_quantizers.34 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.down_proj_weight_quantizers.35 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.down_proj_weight_quantizers.36 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.70e-01 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.down_proj_weight_quantizers.37 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.16e-01 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.down_proj_weight_quantizers.38 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.down_proj_weight_quantizers.39 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.down_proj_weight_quantizers.40 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.down_proj_weight_quantizers.41 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.down_proj_weight_quantizers.42 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.down_proj_weight_quantizers.43 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.down_proj_weight_quantizers.44 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.down_proj_weight_quantizers.45 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.38e-01 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.down_proj_weight_quantizers.46 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.down_proj_weight_quantizers.47 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.down_proj_weight_quantizers.48 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.down_proj_weight_quantizers.49 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.down_proj_weight_quantizers.50 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.35e-02 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.down_proj_weight_quantizers.51 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.down_proj_weight_quantizers.52 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.64e-02 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.down_proj_weight_quantizers.53 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.down_proj_weight_quantizers.54 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.59e-02 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.down_proj_weight_quantizers.55 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.16e-01 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.down_proj_weight_quantizers.56 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.down_proj_weight_quantizers.57 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.down_proj_weight_quantizers.58 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.down_proj_weight_quantizers.59 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.53e-01 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.down_proj_weight_quantizers.60 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.down_proj_weight_quantizers.61 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.down_proj_weight_quantizers.62 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.down_proj_weight_quantizers.63 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.40e-02 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.down_proj_weight_quantizers.64 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.64e-01 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.down_proj_weight_quantizers.65 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.down_proj_weight_quantizers.66 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.down_proj_weight_quantizers.67 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.down_proj_weight_quantizers.68 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.down_proj_weight_quantizers.69 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.down_proj_weight_quantizers.70 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.down_proj_weight_quantizers.71 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.down_proj_weight_quantizers.72 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.40e-02 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.down_proj_weight_quantizers.73 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.down_proj_weight_quantizers.74 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.64e-02 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.down_proj_weight_quantizers.75 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.down_proj_weight_quantizers.76 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.down_proj_weight_quantizers.77 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.down_proj_weight_quantizers.78 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.59e-02 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.down_proj_weight_quantizers.79 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.64e-01 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.down_proj_weight_quantizers.80 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.down_proj_weight_quantizers.81 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.down_proj_weight_quantizers.82 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.down_proj_weight_quantizers.83 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.26e-01 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.down_proj_weight_quantizers.84 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.down_proj_weight_quantizers.85 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.down_proj_weight_quantizers.86 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.down_proj_weight_quantizers.87 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.down_proj_weight_quantizers.88 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.40e-02 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.down_proj_weight_quantizers.89 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.down_proj_weight_quantizers.90 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.down_proj_weight_quantizers.91 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.down_proj_weight_quantizers.92 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.down_proj_weight_quantizers.93 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.down_proj_weight_quantizers.94 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.50e-02 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.down_proj_weight_quantizers.95 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.down_proj_weight_quantizers.96 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.down_proj_weight_quantizers.97 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.54e-02 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.down_proj_weight_quantizers.98 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.64e-02 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.down_proj_weight_quantizers.99 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.down_proj_weight_quantizers.100 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.down_proj_weight_quantizers.101 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.down_proj_weight_quantizers.102 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.28e-01 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.down_proj_weight_quantizers.103 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.down_proj_weight_quantizers.104 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.89e-01 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.down_proj_weight_quantizers.105 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.down_proj_weight_quantizers.106 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.down_proj_weight_quantizers.107 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.45e-02 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.down_proj_weight_quantizers.108 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.59e-02 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.down_proj_weight_quantizers.109 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.50e-02 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.down_proj_weight_quantizers.110 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.down_proj_weight_quantizers.111 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.down_proj_weight_quantizers.112 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.down_proj_weight_quantizers.113 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.down_proj_weight_quantizers.114 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.down_proj_weight_quantizers.115 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.27e-01 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.down_proj_weight_quantizers.116 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.down_proj_weight_quantizers.117 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.down_proj_weight_quantizers.118 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.down_proj_weight_quantizers.119 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.down_proj_weight_quantizers.120 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.54e-02 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.down_proj_weight_quantizers.121 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.down_proj_weight_quantizers.122 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.down_proj_weight_quantizers.123 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.down_proj_weight_quantizers.124 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.down_proj_weight_quantizers.125 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.40e-02 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.down_proj_weight_quantizers.126 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.down_proj_weight_quantizers.127 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.down_proj_weight_quantizers.128 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.down_proj_weight_quantizers.129 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.down_proj_weight_quantizers.130 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.down_proj_weight_quantizers.131 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.down_proj_weight_quantizers.132 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.down_proj_weight_quantizers.133 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.59e-02 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.down_proj_weight_quantizers.134 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.down_proj_weight_quantizers.135 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.down_proj_weight_quantizers.136 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.77e-01 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.down_proj_weight_quantizers.137 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.down_proj_weight_quantizers.138 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.down_proj_weight_quantizers.139 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.down_proj_weight_quantizers.140 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.50e-02 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.down_proj_weight_quantizers.141 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.down_proj_weight_quantizers.142 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.down_proj_weight_quantizers.143 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.down_proj_weight_quantizers.144 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.down_proj_weight_quantizers.145 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.down_proj_weight_quantizers.146 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.down_proj_weight_quantizers.147 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.down_proj_weight_quantizers.148 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.down_proj_weight_quantizers.149 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.27e-01 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.down_proj_weight_quantizers.150 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.down_proj_weight_quantizers.151 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.14e-01 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.down_proj_weight_quantizers.152 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.down_proj_weight_quantizers.153 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.down_proj_weight_quantizers.154 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.down_proj_weight_quantizers.155 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.down_proj_weight_quantizers.156 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.down_proj_weight_quantizers.157 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.40e-02 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.down_proj_weight_quantizers.158 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.50e-02 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.down_proj_weight_quantizers.159 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.down_proj_weight_quantizers.160 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.down_proj_weight_quantizers.161 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.43e-01 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.down_proj_weight_quantizers.162 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.down_proj_weight_quantizers.163 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.down_proj_weight_quantizers.164 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.down_proj_weight_quantizers.165 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.down_proj_weight_quantizers.166 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.38e-01 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.down_proj_weight_quantizers.167 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.down_proj_weight_quantizers.168 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.down_proj_weight_quantizers.169 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.81e-01 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.down_proj_weight_quantizers.170 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.54e-02 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.down_proj_weight_quantizers.171 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.down_proj_weight_quantizers.172 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.70e-01 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.down_proj_weight_quantizers.173 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.48e-01 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.down_proj_weight_quantizers.174 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.down_proj_weight_quantizers.175 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.37e-01 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.down_proj_weight_quantizers.176 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.down_proj_weight_quantizers.177 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.down_proj_weight_quantizers.178 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.down_proj_weight_quantizers.179 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.down_proj_weight_quantizers.180 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.down_proj_weight_quantizers.181 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.down_proj_weight_quantizers.182 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.down_proj_weight_quantizers.183 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.down_proj_weight_quantizers.184 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.down_proj_weight_quantizers.185 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.down_proj_weight_quantizers.186 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.30e-01 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.down_proj_weight_quantizers.187 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.down_proj_weight_quantizers.188 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.82e-01 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.down_proj_weight_quantizers.189 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.down_proj_weight_quantizers.190 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.down_proj_weight_quantizers.191 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.down_proj_weight_quantizers.192 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.down_proj_weight_quantizers.193 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.down_proj_weight_quantizers.194 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.16e-01 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.down_proj_weight_quantizers.195 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.down_proj_weight_quantizers.196 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.down_proj_weight_quantizers.197 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.down_proj_weight_quantizers.198 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.down_proj_weight_quantizers.199 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.72e-01 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.down_proj_weight_quantizers.200 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.down_proj_weight_quantizers.201 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.down_proj_weight_quantizers.202 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.54e-02 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.down_proj_weight_quantizers.203 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.down_proj_weight_quantizers.204 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.down_proj_weight_quantizers.205 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.down_proj_weight_quantizers.206 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.down_proj_weight_quantizers.207 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.43e-01 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.down_proj_weight_quantizers.208 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.down_proj_weight_quantizers.209 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.50e-02 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.down_proj_weight_quantizers.210 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.59e-02 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.down_proj_weight_quantizers.211 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.down_proj_weight_quantizers.212 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.70e-01 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.down_proj_weight_quantizers.213 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.down_proj_weight_quantizers.214 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.down_proj_weight_quantizers.215 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.down_proj_weight_quantizers.216 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.down_proj_weight_quantizers.217 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.down_proj_weight_quantizers.218 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.down_proj_weight_quantizers.219 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.down_proj_weight_quantizers.220 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.down_proj_weight_quantizers.221 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.down_proj_weight_quantizers.222 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.down_proj_weight_quantizers.223 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.down_proj_weight_quantizers.224 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.71e-01 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.down_proj_weight_quantizers.225 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.40e-02 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.down_proj_weight_quantizers.226 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.down_proj_weight_quantizers.227 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.30e-02 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.down_proj_weight_quantizers.228 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.down_proj_weight_quantizers.229 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.63e-01 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.down_proj_weight_quantizers.230 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.down_proj_weight_quantizers.231 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.down_proj_weight_quantizers.232 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.down_proj_weight_quantizers.233 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.64e-02 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.down_proj_weight_quantizers.234 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.51e-01 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.down_proj_weight_quantizers.235 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.down_proj_weight_quantizers.236 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.down_proj_weight_quantizers.237 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.54e-02 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.down_proj_weight_quantizers.238 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.down_proj_weight_quantizers.239 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.down_proj_weight_quantizers.240 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.down_proj_weight_quantizers.241 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.down_proj_weight_quantizers.242 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.down_proj_weight_quantizers.243 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.down_proj_weight_quantizers.244 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.down_proj_weight_quantizers.245 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.40e-02 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.down_proj_weight_quantizers.246 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.down_proj_weight_quantizers.247 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.45e-02 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.down_proj_weight_quantizers.248 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.40e-02 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.down_proj_weight_quantizers.249 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.down_proj_weight_quantizers.250 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.down_proj_weight_quantizers.251 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.59e-02 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.down_proj_weight_quantizers.252 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.down_proj_weight_quantizers.253 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.64e-02 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.down_proj_weight_quantizers.254 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.down_proj_weight_quantizers.255 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.16e-01 calibrator=MaxCalibrator quant) +model.layers.28.mlp.shared_experts.gate_proj.input_quantizer TensorQuantizer(disabled) +model.layers.28.mlp.shared_experts.gate_proj.output_quantizer TensorQuantizer(disabled) +model.layers.28.mlp.shared_experts.gate_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.28.mlp.shared_experts.up_proj.input_quantizer TensorQuantizer(disabled) +model.layers.28.mlp.shared_experts.up_proj.output_quantizer TensorQuantizer(disabled) +model.layers.28.mlp.shared_experts.up_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.28.mlp.shared_experts.down_proj.input_quantizer TensorQuantizer(disabled) +model.layers.28.mlp.shared_experts.down_proj.output_quantizer TensorQuantizer(disabled) +model.layers.28.mlp.shared_experts.down_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.29.self_attn.q_a_proj.input_quantizer TensorQuantizer(disabled) +model.layers.29.self_attn.q_a_proj.output_quantizer TensorQuantizer(disabled) +model.layers.29.self_attn.q_a_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.29.self_attn.q_b_proj.input_quantizer TensorQuantizer(disabled) +model.layers.29.self_attn.q_b_proj.output_quantizer TensorQuantizer(disabled) +model.layers.29.self_attn.q_b_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.29.self_attn.kv_a_proj_with_mqa.input_quantizer TensorQuantizer(disabled) +model.layers.29.self_attn.kv_a_proj_with_mqa.output_quantizer TensorQuantizer(disabled) +model.layers.29.self_attn.kv_a_proj_with_mqa.weight_quantizer TensorQuantizer(disabled) +model.layers.29.self_attn.kv_b_proj.input_quantizer TensorQuantizer(disabled) +model.layers.29.self_attn.kv_b_proj.output_quantizer TensorQuantizer(disabled) +model.layers.29.self_attn.kv_b_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.29.self_attn.o_proj.input_quantizer TensorQuantizer(disabled) +model.layers.29.self_attn.o_proj.output_quantizer TensorQuantizer(disabled) +model.layers.29.self_attn.o_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.29.self_attn.q_bmm_quantizer TensorQuantizer(disabled) +model.layers.29.self_attn.k_bmm_quantizer TensorQuantizer((4, 3) bit fake per-tensor amax=4.48e+02(const) calibrator=MaxCalibrator quant) +model.layers.29.self_attn.v_bmm_quantizer TensorQuantizer((4, 3) bit fake per-tensor amax=4.48e+02(const) calibrator=MaxCalibrator quant) +model.layers.29.self_attn.p_bmm_quantizer TensorQuantizer(disabled) +model.layers.29.mlp.experts.gate_up_proj_input_quantizer TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e+00 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.gate_up_proj_weight_quantizers.0 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.gate_up_proj_weight_quantizers.1 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.gate_up_proj_weight_quantizers.2 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.gate_up_proj_weight_quantizers.3 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.gate_up_proj_weight_quantizers.4 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.gate_up_proj_weight_quantizers.5 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.13e-01 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.gate_up_proj_weight_quantizers.6 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.gate_up_proj_weight_quantizers.7 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.gate_up_proj_weight_quantizers.8 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.gate_up_proj_weight_quantizers.9 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.gate_up_proj_weight_quantizers.10 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.gate_up_proj_weight_quantizers.11 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.gate_up_proj_weight_quantizers.12 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.gate_up_proj_weight_quantizers.13 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.gate_up_proj_weight_quantizers.14 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.gate_up_proj_weight_quantizers.15 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.gate_up_proj_weight_quantizers.16 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.16e-01 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.gate_up_proj_weight_quantizers.17 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.50e-01 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.gate_up_proj_weight_quantizers.18 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.gate_up_proj_weight_quantizers.19 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.29e-01 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.gate_up_proj_weight_quantizers.20 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.gate_up_proj_weight_quantizers.21 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.25e-01 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.gate_up_proj_weight_quantizers.22 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.gate_up_proj_weight_quantizers.23 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.26e-01 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.gate_up_proj_weight_quantizers.24 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.gate_up_proj_weight_quantizers.25 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.gate_up_proj_weight_quantizers.26 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.28e-01 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.gate_up_proj_weight_quantizers.27 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.gate_up_proj_weight_quantizers.28 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.85e-01 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.gate_up_proj_weight_quantizers.29 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.gate_up_proj_weight_quantizers.30 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.91e-01 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.gate_up_proj_weight_quantizers.31 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.gate_up_proj_weight_quantizers.32 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.gate_up_proj_weight_quantizers.33 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.gate_up_proj_weight_quantizers.34 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.68e-01 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.gate_up_proj_weight_quantizers.35 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.53e-01 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.gate_up_proj_weight_quantizers.36 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.gate_up_proj_weight_quantizers.37 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.gate_up_proj_weight_quantizers.38 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.gate_up_proj_weight_quantizers.39 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.gate_up_proj_weight_quantizers.40 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.gate_up_proj_weight_quantizers.41 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.gate_up_proj_weight_quantizers.42 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.19e-01 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.gate_up_proj_weight_quantizers.43 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.gate_up_proj_weight_quantizers.44 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.35e-01 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.gate_up_proj_weight_quantizers.45 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.gate_up_proj_weight_quantizers.46 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.gate_up_proj_weight_quantizers.47 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.gate_up_proj_weight_quantizers.48 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.gate_up_proj_weight_quantizers.49 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.gate_up_proj_weight_quantizers.50 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.gate_up_proj_weight_quantizers.51 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.gate_up_proj_weight_quantizers.52 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.gate_up_proj_weight_quantizers.53 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.gate_up_proj_weight_quantizers.54 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.gate_up_proj_weight_quantizers.55 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.gate_up_proj_weight_quantizers.56 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.26e-01 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.gate_up_proj_weight_quantizers.57 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.gate_up_proj_weight_quantizers.58 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.gate_up_proj_weight_quantizers.59 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.14e-01 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.gate_up_proj_weight_quantizers.60 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.gate_up_proj_weight_quantizers.61 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.gate_up_proj_weight_quantizers.62 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.gate_up_proj_weight_quantizers.63 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.58e-01 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.gate_up_proj_weight_quantizers.64 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.gate_up_proj_weight_quantizers.65 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.18e-01 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.gate_up_proj_weight_quantizers.66 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.gate_up_proj_weight_quantizers.67 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.gate_up_proj_weight_quantizers.68 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.gate_up_proj_weight_quantizers.69 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.gate_up_proj_weight_quantizers.70 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.gate_up_proj_weight_quantizers.71 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.gate_up_proj_weight_quantizers.72 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.15e-01 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.gate_up_proj_weight_quantizers.73 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.gate_up_proj_weight_quantizers.74 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.78e-01 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.gate_up_proj_weight_quantizers.75 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.gate_up_proj_weight_quantizers.76 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.51e-01 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.gate_up_proj_weight_quantizers.77 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.gate_up_proj_weight_quantizers.78 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.gate_up_proj_weight_quantizers.79 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.gate_up_proj_weight_quantizers.80 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=4.32e-01 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.gate_up_proj_weight_quantizers.81 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.gate_up_proj_weight_quantizers.82 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.gate_up_proj_weight_quantizers.83 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.gate_up_proj_weight_quantizers.84 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.43e-01 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.gate_up_proj_weight_quantizers.85 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.gate_up_proj_weight_quantizers.86 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.41e-01 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.gate_up_proj_weight_quantizers.87 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.16e-01 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.gate_up_proj_weight_quantizers.88 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.gate_up_proj_weight_quantizers.89 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.gate_up_proj_weight_quantizers.90 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.gate_up_proj_weight_quantizers.91 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.gate_up_proj_weight_quantizers.92 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.gate_up_proj_weight_quantizers.93 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.16e-01 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.gate_up_proj_weight_quantizers.94 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.gate_up_proj_weight_quantizers.95 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.gate_up_proj_weight_quantizers.96 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.gate_up_proj_weight_quantizers.97 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.46e-01 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.gate_up_proj_weight_quantizers.98 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.gate_up_proj_weight_quantizers.99 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.20e-01 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.gate_up_proj_weight_quantizers.100 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.gate_up_proj_weight_quantizers.101 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.gate_up_proj_weight_quantizers.102 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.37e-01 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.gate_up_proj_weight_quantizers.103 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.gate_up_proj_weight_quantizers.104 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.gate_up_proj_weight_quantizers.105 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.gate_up_proj_weight_quantizers.106 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.22e-01 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.gate_up_proj_weight_quantizers.107 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.17e-01 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.gate_up_proj_weight_quantizers.108 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.gate_up_proj_weight_quantizers.109 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.gate_up_proj_weight_quantizers.110 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.68e-01 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.gate_up_proj_weight_quantizers.111 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.gate_up_proj_weight_quantizers.112 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.70e-01 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.gate_up_proj_weight_quantizers.113 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.70e-01 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.gate_up_proj_weight_quantizers.114 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.gate_up_proj_weight_quantizers.115 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.gate_up_proj_weight_quantizers.116 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.17e-01 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.gate_up_proj_weight_quantizers.117 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.gate_up_proj_weight_quantizers.118 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.gate_up_proj_weight_quantizers.119 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.gate_up_proj_weight_quantizers.120 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.85e-01 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.gate_up_proj_weight_quantizers.121 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.gate_up_proj_weight_quantizers.122 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.gate_up_proj_weight_quantizers.123 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.gate_up_proj_weight_quantizers.124 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.gate_up_proj_weight_quantizers.125 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.gate_up_proj_weight_quantizers.126 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.gate_up_proj_weight_quantizers.127 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.12e-01 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.gate_up_proj_weight_quantizers.128 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.gate_up_proj_weight_quantizers.129 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.29e-01 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.gate_up_proj_weight_quantizers.130 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.gate_up_proj_weight_quantizers.131 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.gate_up_proj_weight_quantizers.132 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.gate_up_proj_weight_quantizers.133 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.gate_up_proj_weight_quantizers.134 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.gate_up_proj_weight_quantizers.135 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.15e-01 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.gate_up_proj_weight_quantizers.136 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.gate_up_proj_weight_quantizers.137 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.gate_up_proj_weight_quantizers.138 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.gate_up_proj_weight_quantizers.139 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.gate_up_proj_weight_quantizers.140 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.gate_up_proj_weight_quantizers.141 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.56e-01 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.gate_up_proj_weight_quantizers.142 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.gate_up_proj_weight_quantizers.143 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.gate_up_proj_weight_quantizers.144 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.22e-01 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.gate_up_proj_weight_quantizers.145 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.gate_up_proj_weight_quantizers.146 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.gate_up_proj_weight_quantizers.147 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.gate_up_proj_weight_quantizers.148 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.gate_up_proj_weight_quantizers.149 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.gate_up_proj_weight_quantizers.150 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.gate_up_proj_weight_quantizers.151 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.gate_up_proj_weight_quantizers.152 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.gate_up_proj_weight_quantizers.153 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.gate_up_proj_weight_quantizers.154 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.gate_up_proj_weight_quantizers.155 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.26e-01 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.gate_up_proj_weight_quantizers.156 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.gate_up_proj_weight_quantizers.157 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.34e-01 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.gate_up_proj_weight_quantizers.158 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.gate_up_proj_weight_quantizers.159 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.gate_up_proj_weight_quantizers.160 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.gate_up_proj_weight_quantizers.161 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.gate_up_proj_weight_quantizers.162 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.23e-01 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.gate_up_proj_weight_quantizers.163 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.71e-01 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.gate_up_proj_weight_quantizers.164 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.gate_up_proj_weight_quantizers.165 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.gate_up_proj_weight_quantizers.166 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.gate_up_proj_weight_quantizers.167 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.44e-01 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.gate_up_proj_weight_quantizers.168 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.gate_up_proj_weight_quantizers.169 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.gate_up_proj_weight_quantizers.170 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.gate_up_proj_weight_quantizers.171 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.14e-01 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.gate_up_proj_weight_quantizers.172 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.gate_up_proj_weight_quantizers.173 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.07e-01 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.gate_up_proj_weight_quantizers.174 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.gate_up_proj_weight_quantizers.175 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.19e-01 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.gate_up_proj_weight_quantizers.176 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.20e-01 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.gate_up_proj_weight_quantizers.177 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.gate_up_proj_weight_quantizers.178 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.gate_up_proj_weight_quantizers.179 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.34e-01 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.gate_up_proj_weight_quantizers.180 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.93e-01 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.gate_up_proj_weight_quantizers.181 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.gate_up_proj_weight_quantizers.182 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.gate_up_proj_weight_quantizers.183 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.gate_up_proj_weight_quantizers.184 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.gate_up_proj_weight_quantizers.185 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.gate_up_proj_weight_quantizers.186 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.gate_up_proj_weight_quantizers.187 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.gate_up_proj_weight_quantizers.188 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.gate_up_proj_weight_quantizers.189 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.gate_up_proj_weight_quantizers.190 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.gate_up_proj_weight_quantizers.191 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.gate_up_proj_weight_quantizers.192 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.35e-01 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.gate_up_proj_weight_quantizers.193 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.gate_up_proj_weight_quantizers.194 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.gate_up_proj_weight_quantizers.195 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.64e-01 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.gate_up_proj_weight_quantizers.196 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.31e-01 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.gate_up_proj_weight_quantizers.197 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.54e-01 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.gate_up_proj_weight_quantizers.198 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.14e-01 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.gate_up_proj_weight_quantizers.199 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.gate_up_proj_weight_quantizers.200 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.30e-01 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.gate_up_proj_weight_quantizers.201 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.33e-01 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.gate_up_proj_weight_quantizers.202 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.gate_up_proj_weight_quantizers.203 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.59e-01 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.gate_up_proj_weight_quantizers.204 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.18e-01 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.gate_up_proj_weight_quantizers.205 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.gate_up_proj_weight_quantizers.206 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.gate_up_proj_weight_quantizers.207 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.gate_up_proj_weight_quantizers.208 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.32e-01 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.gate_up_proj_weight_quantizers.209 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.39e-01 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.gate_up_proj_weight_quantizers.210 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.44e-01 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.gate_up_proj_weight_quantizers.211 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.gate_up_proj_weight_quantizers.212 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.gate_up_proj_weight_quantizers.213 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.79e-01 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.gate_up_proj_weight_quantizers.214 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.gate_up_proj_weight_quantizers.215 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.gate_up_proj_weight_quantizers.216 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.69e-01 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.gate_up_proj_weight_quantizers.217 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.18e-01 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.gate_up_proj_weight_quantizers.218 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.gate_up_proj_weight_quantizers.219 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.gate_up_proj_weight_quantizers.220 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.gate_up_proj_weight_quantizers.221 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.gate_up_proj_weight_quantizers.222 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.gate_up_proj_weight_quantizers.223 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.gate_up_proj_weight_quantizers.224 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.gate_up_proj_weight_quantizers.225 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.36e-01 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.gate_up_proj_weight_quantizers.226 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.gate_up_proj_weight_quantizers.227 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.gate_up_proj_weight_quantizers.228 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.gate_up_proj_weight_quantizers.229 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.gate_up_proj_weight_quantizers.230 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.95e-01 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.gate_up_proj_weight_quantizers.231 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.gate_up_proj_weight_quantizers.232 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.gate_up_proj_weight_quantizers.233 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.gate_up_proj_weight_quantizers.234 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.gate_up_proj_weight_quantizers.235 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.26e-01 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.gate_up_proj_weight_quantizers.236 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.gate_up_proj_weight_quantizers.237 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.gate_up_proj_weight_quantizers.238 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.gate_up_proj_weight_quantizers.239 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.gate_up_proj_weight_quantizers.240 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.gate_up_proj_weight_quantizers.241 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.gate_up_proj_weight_quantizers.242 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.54e-01 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.gate_up_proj_weight_quantizers.243 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.gate_up_proj_weight_quantizers.244 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.gate_up_proj_weight_quantizers.245 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.gate_up_proj_weight_quantizers.246 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.gate_up_proj_weight_quantizers.247 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.gate_up_proj_weight_quantizers.248 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.gate_up_proj_weight_quantizers.249 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.gate_up_proj_weight_quantizers.250 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.gate_up_proj_weight_quantizers.251 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.gate_up_proj_weight_quantizers.252 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.gate_up_proj_weight_quantizers.253 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.gate_up_proj_weight_quantizers.254 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.gate_up_proj_weight_quantizers.255 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.97e-01 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.down_proj_input_quantizer TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.50e+01 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.down_proj_weight_quantizers.0 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.down_proj_weight_quantizers.1 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.down_proj_weight_quantizers.2 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.down_proj_weight_quantizers.3 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.down_proj_weight_quantizers.4 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.down_proj_weight_quantizers.5 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.47e-01 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.down_proj_weight_quantizers.6 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.down_proj_weight_quantizers.7 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.down_proj_weight_quantizers.8 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.59e-02 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.down_proj_weight_quantizers.9 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.59e-02 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.down_proj_weight_quantizers.10 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.down_proj_weight_quantizers.11 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.down_proj_weight_quantizers.12 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.64e-02 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.down_proj_weight_quantizers.13 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.down_proj_weight_quantizers.14 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.57e-01 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.down_proj_weight_quantizers.15 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.down_proj_weight_quantizers.16 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.down_proj_weight_quantizers.17 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.90e-01 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.down_proj_weight_quantizers.18 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.down_proj_weight_quantizers.19 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.68e-01 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.down_proj_weight_quantizers.20 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.50e-02 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.down_proj_weight_quantizers.21 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.16e-01 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.down_proj_weight_quantizers.22 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.down_proj_weight_quantizers.23 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.12e-01 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.down_proj_weight_quantizers.24 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.down_proj_weight_quantizers.25 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.down_proj_weight_quantizers.26 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.66e-01 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.down_proj_weight_quantizers.27 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.down_proj_weight_quantizers.28 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.48e-01 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.down_proj_weight_quantizers.29 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.down_proj_weight_quantizers.30 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.20e-01 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.down_proj_weight_quantizers.31 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.down_proj_weight_quantizers.32 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.down_proj_weight_quantizers.33 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.30e-02 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.down_proj_weight_quantizers.34 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.61e-01 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.down_proj_weight_quantizers.35 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.36e-01 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.down_proj_weight_quantizers.36 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.64e-02 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.down_proj_weight_quantizers.37 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.59e-02 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.down_proj_weight_quantizers.38 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.down_proj_weight_quantizers.39 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.down_proj_weight_quantizers.40 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.down_proj_weight_quantizers.41 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.down_proj_weight_quantizers.42 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.down_proj_weight_quantizers.43 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.down_proj_weight_quantizers.44 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.64e-02 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.down_proj_weight_quantizers.45 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.down_proj_weight_quantizers.46 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.down_proj_weight_quantizers.47 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.down_proj_weight_quantizers.48 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.down_proj_weight_quantizers.49 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.64e-02 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.down_proj_weight_quantizers.50 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.down_proj_weight_quantizers.51 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.down_proj_weight_quantizers.52 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.down_proj_weight_quantizers.53 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.25e-01 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.down_proj_weight_quantizers.54 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.down_proj_weight_quantizers.55 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.down_proj_weight_quantizers.56 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.down_proj_weight_quantizers.57 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.50e-02 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.down_proj_weight_quantizers.58 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.down_proj_weight_quantizers.59 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.36e-01 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.down_proj_weight_quantizers.60 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.down_proj_weight_quantizers.61 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.down_proj_weight_quantizers.62 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.30e-02 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.down_proj_weight_quantizers.63 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.21e-01 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.down_proj_weight_quantizers.64 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.down_proj_weight_quantizers.65 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.down_proj_weight_quantizers.66 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.down_proj_weight_quantizers.67 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.down_proj_weight_quantizers.68 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.down_proj_weight_quantizers.69 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.64e-02 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.down_proj_weight_quantizers.70 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.down_proj_weight_quantizers.71 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.down_proj_weight_quantizers.72 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.77e-01 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.down_proj_weight_quantizers.73 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.28e-01 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.down_proj_weight_quantizers.74 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.down_proj_weight_quantizers.75 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.down_proj_weight_quantizers.76 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.59e-01 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.down_proj_weight_quantizers.77 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.16e-01 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.down_proj_weight_quantizers.78 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.59e-02 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.down_proj_weight_quantizers.79 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.down_proj_weight_quantizers.80 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.44e-01 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.down_proj_weight_quantizers.81 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.down_proj_weight_quantizers.82 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.down_proj_weight_quantizers.83 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.down_proj_weight_quantizers.84 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.down_proj_weight_quantizers.85 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.54e-02 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.down_proj_weight_quantizers.86 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.38e-01 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.down_proj_weight_quantizers.87 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.45e-01 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.down_proj_weight_quantizers.88 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.down_proj_weight_quantizers.89 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.down_proj_weight_quantizers.90 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.down_proj_weight_quantizers.91 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.down_proj_weight_quantizers.92 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.50e-02 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.down_proj_weight_quantizers.93 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.38e-01 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.down_proj_weight_quantizers.94 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.down_proj_weight_quantizers.95 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.35e-01 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.down_proj_weight_quantizers.96 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.down_proj_weight_quantizers.97 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.down_proj_weight_quantizers.98 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.down_proj_weight_quantizers.99 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.down_proj_weight_quantizers.100 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.64e-02 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.down_proj_weight_quantizers.101 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.down_proj_weight_quantizers.102 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.35e-01 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.down_proj_weight_quantizers.103 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.down_proj_weight_quantizers.104 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.30e-01 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.down_proj_weight_quantizers.105 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.down_proj_weight_quantizers.106 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.down_proj_weight_quantizers.107 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.down_proj_weight_quantizers.108 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.down_proj_weight_quantizers.109 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.down_proj_weight_quantizers.110 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.32e-01 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.down_proj_weight_quantizers.111 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.54e-02 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.down_proj_weight_quantizers.112 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.down_proj_weight_quantizers.113 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.27e-01 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.down_proj_weight_quantizers.114 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.38e-01 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.down_proj_weight_quantizers.115 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.down_proj_weight_quantizers.116 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.down_proj_weight_quantizers.117 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.down_proj_weight_quantizers.118 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.down_proj_weight_quantizers.119 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.down_proj_weight_quantizers.120 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.56e-01 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.down_proj_weight_quantizers.121 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.74e-01 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.down_proj_weight_quantizers.122 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.down_proj_weight_quantizers.123 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.down_proj_weight_quantizers.124 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.down_proj_weight_quantizers.125 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.40e-02 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.down_proj_weight_quantizers.126 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.down_proj_weight_quantizers.127 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.34e-01 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.down_proj_weight_quantizers.128 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.59e-02 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.down_proj_weight_quantizers.129 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.down_proj_weight_quantizers.130 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.down_proj_weight_quantizers.131 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.down_proj_weight_quantizers.132 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.28e-01 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.down_proj_weight_quantizers.133 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.down_proj_weight_quantizers.134 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.64e-02 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.down_proj_weight_quantizers.135 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.23e-01 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.down_proj_weight_quantizers.136 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.down_proj_weight_quantizers.137 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.down_proj_weight_quantizers.138 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.down_proj_weight_quantizers.139 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.down_proj_weight_quantizers.140 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.down_proj_weight_quantizers.141 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.67e-01 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.down_proj_weight_quantizers.142 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.down_proj_weight_quantizers.143 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.down_proj_weight_quantizers.144 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.down_proj_weight_quantizers.145 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.down_proj_weight_quantizers.146 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.down_proj_weight_quantizers.147 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.40e-02 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.down_proj_weight_quantizers.148 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.down_proj_weight_quantizers.149 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.down_proj_weight_quantizers.150 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.down_proj_weight_quantizers.151 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.54e-02 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.down_proj_weight_quantizers.152 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.down_proj_weight_quantizers.153 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.down_proj_weight_quantizers.154 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.down_proj_weight_quantizers.155 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.18e-01 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.down_proj_weight_quantizers.156 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.down_proj_weight_quantizers.157 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.95e-01 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.down_proj_weight_quantizers.158 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.36e-01 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.down_proj_weight_quantizers.159 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.down_proj_weight_quantizers.160 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.down_proj_weight_quantizers.161 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.38e-01 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.down_proj_weight_quantizers.162 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.60e-01 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.down_proj_weight_quantizers.163 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.49e-01 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.down_proj_weight_quantizers.164 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.64e-02 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.down_proj_weight_quantizers.165 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.down_proj_weight_quantizers.166 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.down_proj_weight_quantizers.167 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.down_proj_weight_quantizers.168 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.down_proj_weight_quantizers.169 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.50e-02 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.down_proj_weight_quantizers.170 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.down_proj_weight_quantizers.171 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.40e-02 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.down_proj_weight_quantizers.172 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.down_proj_weight_quantizers.173 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.97e-01 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.down_proj_weight_quantizers.174 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.down_proj_weight_quantizers.175 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.88e-01 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.down_proj_weight_quantizers.176 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.15e-01 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.down_proj_weight_quantizers.177 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.down_proj_weight_quantizers.178 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.50e-02 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.down_proj_weight_quantizers.179 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.35e-02 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.down_proj_weight_quantizers.180 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.down_proj_weight_quantizers.181 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.59e-02 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.down_proj_weight_quantizers.182 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.19e-01 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.down_proj_weight_quantizers.183 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.40e-02 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.down_proj_weight_quantizers.184 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.down_proj_weight_quantizers.185 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.down_proj_weight_quantizers.186 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.64e-02 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.down_proj_weight_quantizers.187 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.30e-01 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.down_proj_weight_quantizers.188 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.down_proj_weight_quantizers.189 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.down_proj_weight_quantizers.190 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.down_proj_weight_quantizers.191 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.50e-02 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.down_proj_weight_quantizers.192 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.77e-01 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.down_proj_weight_quantizers.193 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.down_proj_weight_quantizers.194 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.down_proj_weight_quantizers.195 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.95e-01 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.down_proj_weight_quantizers.196 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.23e-01 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.down_proj_weight_quantizers.197 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.70e-01 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.down_proj_weight_quantizers.198 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.26e-01 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.down_proj_weight_quantizers.199 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.down_proj_weight_quantizers.200 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.25e-02 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.down_proj_weight_quantizers.201 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.71e-01 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.down_proj_weight_quantizers.202 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.down_proj_weight_quantizers.203 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.down_proj_weight_quantizers.204 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.down_proj_weight_quantizers.205 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.down_proj_weight_quantizers.206 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.down_proj_weight_quantizers.207 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.down_proj_weight_quantizers.208 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.64e-02 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.down_proj_weight_quantizers.209 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.41e-01 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.down_proj_weight_quantizers.210 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.64e-02 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.down_proj_weight_quantizers.211 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.down_proj_weight_quantizers.212 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.down_proj_weight_quantizers.213 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.53e-01 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.down_proj_weight_quantizers.214 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.down_proj_weight_quantizers.215 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.down_proj_weight_quantizers.216 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.88e-01 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.down_proj_weight_quantizers.217 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.50e-02 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.down_proj_weight_quantizers.218 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.down_proj_weight_quantizers.219 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.45e-02 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.down_proj_weight_quantizers.220 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.35e-02 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.down_proj_weight_quantizers.221 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.down_proj_weight_quantizers.222 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.down_proj_weight_quantizers.223 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.down_proj_weight_quantizers.224 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.down_proj_weight_quantizers.225 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.14e-01 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.down_proj_weight_quantizers.226 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.down_proj_weight_quantizers.227 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.down_proj_weight_quantizers.228 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.down_proj_weight_quantizers.229 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.down_proj_weight_quantizers.230 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.76e-01 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.down_proj_weight_quantizers.231 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.down_proj_weight_quantizers.232 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.down_proj_weight_quantizers.233 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.down_proj_weight_quantizers.234 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.down_proj_weight_quantizers.235 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.47e-01 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.down_proj_weight_quantizers.236 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.down_proj_weight_quantizers.237 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.64e-02 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.down_proj_weight_quantizers.238 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.down_proj_weight_quantizers.239 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.59e-02 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.down_proj_weight_quantizers.240 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.20e-02 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.down_proj_weight_quantizers.241 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.down_proj_weight_quantizers.242 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.59e-02 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.down_proj_weight_quantizers.243 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.21e-01 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.down_proj_weight_quantizers.244 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.down_proj_weight_quantizers.245 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.down_proj_weight_quantizers.246 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.down_proj_weight_quantizers.247 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.64e-02 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.down_proj_weight_quantizers.248 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.50e-02 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.down_proj_weight_quantizers.249 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.down_proj_weight_quantizers.250 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.down_proj_weight_quantizers.251 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.50e-02 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.down_proj_weight_quantizers.252 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.down_proj_weight_quantizers.253 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.down_proj_weight_quantizers.254 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.64e-02 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.down_proj_weight_quantizers.255 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.29.mlp.shared_experts.gate_proj.input_quantizer TensorQuantizer(disabled) +model.layers.29.mlp.shared_experts.gate_proj.output_quantizer TensorQuantizer(disabled) +model.layers.29.mlp.shared_experts.gate_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.29.mlp.shared_experts.up_proj.input_quantizer TensorQuantizer(disabled) +model.layers.29.mlp.shared_experts.up_proj.output_quantizer TensorQuantizer(disabled) +model.layers.29.mlp.shared_experts.up_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.29.mlp.shared_experts.down_proj.input_quantizer TensorQuantizer(disabled) +model.layers.29.mlp.shared_experts.down_proj.output_quantizer TensorQuantizer(disabled) +model.layers.29.mlp.shared_experts.down_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.30.self_attn.q_a_proj.input_quantizer TensorQuantizer(disabled) +model.layers.30.self_attn.q_a_proj.output_quantizer TensorQuantizer(disabled) +model.layers.30.self_attn.q_a_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.30.self_attn.q_b_proj.input_quantizer TensorQuantizer(disabled) +model.layers.30.self_attn.q_b_proj.output_quantizer TensorQuantizer(disabled) +model.layers.30.self_attn.q_b_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.30.self_attn.kv_a_proj_with_mqa.input_quantizer TensorQuantizer(disabled) +model.layers.30.self_attn.kv_a_proj_with_mqa.output_quantizer TensorQuantizer(disabled) +model.layers.30.self_attn.kv_a_proj_with_mqa.weight_quantizer TensorQuantizer(disabled) +model.layers.30.self_attn.kv_b_proj.input_quantizer TensorQuantizer(disabled) +model.layers.30.self_attn.kv_b_proj.output_quantizer TensorQuantizer(disabled) +model.layers.30.self_attn.kv_b_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.30.self_attn.o_proj.input_quantizer TensorQuantizer(disabled) +model.layers.30.self_attn.o_proj.output_quantizer TensorQuantizer(disabled) +model.layers.30.self_attn.o_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.30.self_attn.indexer.wq_b.input_quantizer TensorQuantizer(disabled) +model.layers.30.self_attn.indexer.wq_b.output_quantizer TensorQuantizer(disabled) +model.layers.30.self_attn.indexer.wq_b.weight_quantizer TensorQuantizer(disabled) +model.layers.30.self_attn.indexer.wk.input_quantizer TensorQuantizer(disabled) +model.layers.30.self_attn.indexer.wk.output_quantizer TensorQuantizer(disabled) +model.layers.30.self_attn.indexer.wk.weight_quantizer TensorQuantizer(disabled) +model.layers.30.self_attn.indexer.k_norm.input_quantizer TensorQuantizer(disabled) +model.layers.30.self_attn.indexer.k_norm.output_quantizer TensorQuantizer(disabled) +model.layers.30.self_attn.indexer.weights_proj.input_quantizer TensorQuantizer(disabled) +model.layers.30.self_attn.indexer.weights_proj.output_quantizer TensorQuantizer(disabled) +model.layers.30.self_attn.indexer.weights_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.30.self_attn.q_bmm_quantizer TensorQuantizer(disabled) +model.layers.30.self_attn.k_bmm_quantizer TensorQuantizer((4, 3) bit fake per-tensor amax=4.48e+02(const) calibrator=MaxCalibrator quant) +model.layers.30.self_attn.v_bmm_quantizer TensorQuantizer((4, 3) bit fake per-tensor amax=4.48e+02(const) calibrator=MaxCalibrator quant) +model.layers.30.self_attn.p_bmm_quantizer TensorQuantizer(disabled) +model.layers.30.mlp.experts.gate_up_proj_input_quantizer TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e+00 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.gate_up_proj_weight_quantizers.0 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.gate_up_proj_weight_quantizers.1 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.gate_up_proj_weight_quantizers.2 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.13e-01 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.gate_up_proj_weight_quantizers.3 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.67e-01 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.gate_up_proj_weight_quantizers.4 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.gate_up_proj_weight_quantizers.5 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.gate_up_proj_weight_quantizers.6 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.gate_up_proj_weight_quantizers.7 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.gate_up_proj_weight_quantizers.8 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.gate_up_proj_weight_quantizers.9 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.gate_up_proj_weight_quantizers.10 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.gate_up_proj_weight_quantizers.11 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.68e-01 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.gate_up_proj_weight_quantizers.12 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.gate_up_proj_weight_quantizers.13 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.gate_up_proj_weight_quantizers.14 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.gate_up_proj_weight_quantizers.15 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.gate_up_proj_weight_quantizers.16 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.17e-01 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.gate_up_proj_weight_quantizers.17 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.gate_up_proj_weight_quantizers.18 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.gate_up_proj_weight_quantizers.19 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.gate_up_proj_weight_quantizers.20 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.47e-01 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.gate_up_proj_weight_quantizers.21 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.15e-01 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.gate_up_proj_weight_quantizers.22 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.gate_up_proj_weight_quantizers.23 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.gate_up_proj_weight_quantizers.24 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.gate_up_proj_weight_quantizers.25 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.gate_up_proj_weight_quantizers.26 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.gate_up_proj_weight_quantizers.27 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.gate_up_proj_weight_quantizers.28 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.gate_up_proj_weight_quantizers.29 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.gate_up_proj_weight_quantizers.30 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.gate_up_proj_weight_quantizers.31 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.gate_up_proj_weight_quantizers.32 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.gate_up_proj_weight_quantizers.33 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.gate_up_proj_weight_quantizers.34 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.19e-01 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.gate_up_proj_weight_quantizers.35 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.gate_up_proj_weight_quantizers.36 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.gate_up_proj_weight_quantizers.37 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.gate_up_proj_weight_quantizers.38 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.14e-01 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.gate_up_proj_weight_quantizers.39 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=6.25e-01 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.gate_up_proj_weight_quantizers.40 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.gate_up_proj_weight_quantizers.41 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.gate_up_proj_weight_quantizers.42 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.gate_up_proj_weight_quantizers.43 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.gate_up_proj_weight_quantizers.44 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.gate_up_proj_weight_quantizers.45 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.gate_up_proj_weight_quantizers.46 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.gate_up_proj_weight_quantizers.47 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.21e-01 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.gate_up_proj_weight_quantizers.48 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.gate_up_proj_weight_quantizers.49 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.gate_up_proj_weight_quantizers.50 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.gate_up_proj_weight_quantizers.51 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.gate_up_proj_weight_quantizers.52 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.gate_up_proj_weight_quantizers.53 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.gate_up_proj_weight_quantizers.54 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.gate_up_proj_weight_quantizers.55 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.gate_up_proj_weight_quantizers.56 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.gate_up_proj_weight_quantizers.57 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.gate_up_proj_weight_quantizers.58 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.gate_up_proj_weight_quantizers.59 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.38e-01 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.gate_up_proj_weight_quantizers.60 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.gate_up_proj_weight_quantizers.61 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.gate_up_proj_weight_quantizers.62 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.gate_up_proj_weight_quantizers.63 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.gate_up_proj_weight_quantizers.64 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.gate_up_proj_weight_quantizers.65 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.16e-01 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.gate_up_proj_weight_quantizers.66 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.gate_up_proj_weight_quantizers.67 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.gate_up_proj_weight_quantizers.68 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.gate_up_proj_weight_quantizers.69 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.gate_up_proj_weight_quantizers.70 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.gate_up_proj_weight_quantizers.71 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.52e-01 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.gate_up_proj_weight_quantizers.72 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.33e-01 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.gate_up_proj_weight_quantizers.73 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.gate_up_proj_weight_quantizers.74 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.25e-01 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.gate_up_proj_weight_quantizers.75 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.gate_up_proj_weight_quantizers.76 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.gate_up_proj_weight_quantizers.77 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.gate_up_proj_weight_quantizers.78 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.34e-01 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.gate_up_proj_weight_quantizers.79 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.gate_up_proj_weight_quantizers.80 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.gate_up_proj_weight_quantizers.81 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.gate_up_proj_weight_quantizers.82 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.gate_up_proj_weight_quantizers.83 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.gate_up_proj_weight_quantizers.84 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.gate_up_proj_weight_quantizers.85 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.76e-01 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.gate_up_proj_weight_quantizers.86 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.63e-01 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.gate_up_proj_weight_quantizers.87 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.gate_up_proj_weight_quantizers.88 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.61e-01 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.gate_up_proj_weight_quantizers.89 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.20e-01 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.gate_up_proj_weight_quantizers.90 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.21e-01 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.gate_up_proj_weight_quantizers.91 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.66e-01 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.gate_up_proj_weight_quantizers.92 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.88e-01 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.gate_up_proj_weight_quantizers.93 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.gate_up_proj_weight_quantizers.94 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.gate_up_proj_weight_quantizers.95 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.gate_up_proj_weight_quantizers.96 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.gate_up_proj_weight_quantizers.97 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.gate_up_proj_weight_quantizers.98 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.20e-01 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.gate_up_proj_weight_quantizers.99 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.gate_up_proj_weight_quantizers.100 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.68e-01 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.gate_up_proj_weight_quantizers.101 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.gate_up_proj_weight_quantizers.102 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.gate_up_proj_weight_quantizers.103 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.gate_up_proj_weight_quantizers.104 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.gate_up_proj_weight_quantizers.105 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.gate_up_proj_weight_quantizers.106 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.29e-01 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.gate_up_proj_weight_quantizers.107 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.gate_up_proj_weight_quantizers.108 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.gate_up_proj_weight_quantizers.109 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.gate_up_proj_weight_quantizers.110 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.gate_up_proj_weight_quantizers.111 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.gate_up_proj_weight_quantizers.112 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.gate_up_proj_weight_quantizers.113 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.gate_up_proj_weight_quantizers.114 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.gate_up_proj_weight_quantizers.115 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.gate_up_proj_weight_quantizers.116 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.gate_up_proj_weight_quantizers.117 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.gate_up_proj_weight_quantizers.118 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.15e-01 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.gate_up_proj_weight_quantizers.119 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.62e-01 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.gate_up_proj_weight_quantizers.120 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.gate_up_proj_weight_quantizers.121 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.gate_up_proj_weight_quantizers.122 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.gate_up_proj_weight_quantizers.123 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.gate_up_proj_weight_quantizers.124 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.gate_up_proj_weight_quantizers.125 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.gate_up_proj_weight_quantizers.126 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.gate_up_proj_weight_quantizers.127 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.gate_up_proj_weight_quantizers.128 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.gate_up_proj_weight_quantizers.129 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.gate_up_proj_weight_quantizers.130 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.gate_up_proj_weight_quantizers.131 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.gate_up_proj_weight_quantizers.132 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.gate_up_proj_weight_quantizers.133 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.gate_up_proj_weight_quantizers.134 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.gate_up_proj_weight_quantizers.135 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.25e-01 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.gate_up_proj_weight_quantizers.136 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.40e-01 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.gate_up_proj_weight_quantizers.137 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.gate_up_proj_weight_quantizers.138 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.gate_up_proj_weight_quantizers.139 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.gate_up_proj_weight_quantizers.140 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.gate_up_proj_weight_quantizers.141 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.gate_up_proj_weight_quantizers.142 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.gate_up_proj_weight_quantizers.143 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.gate_up_proj_weight_quantizers.144 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.gate_up_proj_weight_quantizers.145 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.gate_up_proj_weight_quantizers.146 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.gate_up_proj_weight_quantizers.147 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.gate_up_proj_weight_quantizers.148 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.gate_up_proj_weight_quantizers.149 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.gate_up_proj_weight_quantizers.150 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.gate_up_proj_weight_quantizers.151 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.gate_up_proj_weight_quantizers.152 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.gate_up_proj_weight_quantizers.153 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.gate_up_proj_weight_quantizers.154 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.gate_up_proj_weight_quantizers.155 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.62e-01 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.gate_up_proj_weight_quantizers.156 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.23e-01 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.gate_up_proj_weight_quantizers.157 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.gate_up_proj_weight_quantizers.158 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.gate_up_proj_weight_quantizers.159 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.gate_up_proj_weight_quantizers.160 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.72e-01 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.gate_up_proj_weight_quantizers.161 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.gate_up_proj_weight_quantizers.162 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.gate_up_proj_weight_quantizers.163 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.gate_up_proj_weight_quantizers.164 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.20e-01 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.gate_up_proj_weight_quantizers.165 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.46e-01 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.gate_up_proj_weight_quantizers.166 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.gate_up_proj_weight_quantizers.167 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.gate_up_proj_weight_quantizers.168 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.gate_up_proj_weight_quantizers.169 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.35e-01 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.gate_up_proj_weight_quantizers.170 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.gate_up_proj_weight_quantizers.171 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.gate_up_proj_weight_quantizers.172 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.gate_up_proj_weight_quantizers.173 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.15e-01 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.gate_up_proj_weight_quantizers.174 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.23e-01 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.gate_up_proj_weight_quantizers.175 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.gate_up_proj_weight_quantizers.176 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.gate_up_proj_weight_quantizers.177 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.31e-01 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.gate_up_proj_weight_quantizers.178 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.gate_up_proj_weight_quantizers.179 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.16e-01 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.gate_up_proj_weight_quantizers.180 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.40e-01 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.gate_up_proj_weight_quantizers.181 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.gate_up_proj_weight_quantizers.182 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.gate_up_proj_weight_quantizers.183 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.gate_up_proj_weight_quantizers.184 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.gate_up_proj_weight_quantizers.185 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.gate_up_proj_weight_quantizers.186 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.60e-01 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.gate_up_proj_weight_quantizers.187 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.gate_up_proj_weight_quantizers.188 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.88e-01 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.gate_up_proj_weight_quantizers.189 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.gate_up_proj_weight_quantizers.190 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.gate_up_proj_weight_quantizers.191 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.gate_up_proj_weight_quantizers.192 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.gate_up_proj_weight_quantizers.193 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.26e-01 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.gate_up_proj_weight_quantizers.194 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.gate_up_proj_weight_quantizers.195 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.83e-01 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.gate_up_proj_weight_quantizers.196 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.gate_up_proj_weight_quantizers.197 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.gate_up_proj_weight_quantizers.198 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.gate_up_proj_weight_quantizers.199 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.90e-01 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.gate_up_proj_weight_quantizers.200 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.gate_up_proj_weight_quantizers.201 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.37e-01 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.gate_up_proj_weight_quantizers.202 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.gate_up_proj_weight_quantizers.203 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.gate_up_proj_weight_quantizers.204 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.gate_up_proj_weight_quantizers.205 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.14e-01 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.gate_up_proj_weight_quantizers.206 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.88e-01 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.gate_up_proj_weight_quantizers.207 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.36e-01 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.gate_up_proj_weight_quantizers.208 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.gate_up_proj_weight_quantizers.209 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.gate_up_proj_weight_quantizers.210 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.gate_up_proj_weight_quantizers.211 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.gate_up_proj_weight_quantizers.212 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.gate_up_proj_weight_quantizers.213 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.gate_up_proj_weight_quantizers.214 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.82e-01 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.gate_up_proj_weight_quantizers.215 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.gate_up_proj_weight_quantizers.216 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.23e-01 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.gate_up_proj_weight_quantizers.217 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.25e-01 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.gate_up_proj_weight_quantizers.218 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.gate_up_proj_weight_quantizers.219 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.gate_up_proj_weight_quantizers.220 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.39e-01 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.gate_up_proj_weight_quantizers.221 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.22e-01 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.gate_up_proj_weight_quantizers.222 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.gate_up_proj_weight_quantizers.223 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.gate_up_proj_weight_quantizers.224 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.gate_up_proj_weight_quantizers.225 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.gate_up_proj_weight_quantizers.226 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.18e-01 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.gate_up_proj_weight_quantizers.227 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.gate_up_proj_weight_quantizers.228 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.gate_up_proj_weight_quantizers.229 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.20e-01 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.gate_up_proj_weight_quantizers.230 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.69e-01 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.gate_up_proj_weight_quantizers.231 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.gate_up_proj_weight_quantizers.232 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.gate_up_proj_weight_quantizers.233 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.gate_up_proj_weight_quantizers.234 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.gate_up_proj_weight_quantizers.235 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.gate_up_proj_weight_quantizers.236 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.85e-01 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.gate_up_proj_weight_quantizers.237 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.gate_up_proj_weight_quantizers.238 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.gate_up_proj_weight_quantizers.239 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.gate_up_proj_weight_quantizers.240 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.gate_up_proj_weight_quantizers.241 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.gate_up_proj_weight_quantizers.242 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.gate_up_proj_weight_quantizers.243 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.gate_up_proj_weight_quantizers.244 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.gate_up_proj_weight_quantizers.245 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.gate_up_proj_weight_quantizers.246 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.48e-01 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.gate_up_proj_weight_quantizers.247 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.18e-01 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.gate_up_proj_weight_quantizers.248 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.gate_up_proj_weight_quantizers.249 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.gate_up_proj_weight_quantizers.250 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.30e-01 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.gate_up_proj_weight_quantizers.251 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.gate_up_proj_weight_quantizers.252 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.gate_up_proj_weight_quantizers.253 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.gate_up_proj_weight_quantizers.254 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.14e-01 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.gate_up_proj_weight_quantizers.255 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.down_proj_input_quantizer TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=5.20e+01 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.down_proj_weight_quantizers.0 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.50e-01 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.down_proj_weight_quantizers.1 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.down_proj_weight_quantizers.2 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.down_proj_weight_quantizers.3 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.10e-01 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.down_proj_weight_quantizers.4 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.down_proj_weight_quantizers.5 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.down_proj_weight_quantizers.6 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.down_proj_weight_quantizers.7 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.64e-02 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.down_proj_weight_quantizers.8 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.50e-02 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.down_proj_weight_quantizers.9 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.50e-02 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.down_proj_weight_quantizers.10 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.down_proj_weight_quantizers.11 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.19e-01 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.down_proj_weight_quantizers.12 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.down_proj_weight_quantizers.13 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.down_proj_weight_quantizers.14 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.down_proj_weight_quantizers.15 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.64e-02 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.down_proj_weight_quantizers.16 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.down_proj_weight_quantizers.17 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.54e-02 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.down_proj_weight_quantizers.18 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.down_proj_weight_quantizers.19 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.down_proj_weight_quantizers.20 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.64e-01 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.down_proj_weight_quantizers.21 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.40e-02 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.down_proj_weight_quantizers.22 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.14e-01 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.down_proj_weight_quantizers.23 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.down_proj_weight_quantizers.24 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.down_proj_weight_quantizers.25 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.down_proj_weight_quantizers.26 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.30e-02 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.down_proj_weight_quantizers.27 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.down_proj_weight_quantizers.28 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.45e-02 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.down_proj_weight_quantizers.29 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.down_proj_weight_quantizers.30 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.50e-02 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.down_proj_weight_quantizers.31 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.down_proj_weight_quantizers.32 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.down_proj_weight_quantizers.33 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.down_proj_weight_quantizers.34 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.down_proj_weight_quantizers.35 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.down_proj_weight_quantizers.36 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.down_proj_weight_quantizers.37 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.down_proj_weight_quantizers.38 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.down_proj_weight_quantizers.39 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.54e-01 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.down_proj_weight_quantizers.40 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.64e-02 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.down_proj_weight_quantizers.41 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.down_proj_weight_quantizers.42 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.down_proj_weight_quantizers.43 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.down_proj_weight_quantizers.44 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.down_proj_weight_quantizers.45 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.50e-02 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.down_proj_weight_quantizers.46 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.50e-02 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.down_proj_weight_quantizers.47 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.down_proj_weight_quantizers.48 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.59e-02 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.down_proj_weight_quantizers.49 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.down_proj_weight_quantizers.50 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.down_proj_weight_quantizers.51 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.down_proj_weight_quantizers.52 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.down_proj_weight_quantizers.53 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.down_proj_weight_quantizers.54 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.down_proj_weight_quantizers.55 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.down_proj_weight_quantizers.56 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.down_proj_weight_quantizers.57 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.down_proj_weight_quantizers.58 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.54e-02 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.down_proj_weight_quantizers.59 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.down_proj_weight_quantizers.60 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.down_proj_weight_quantizers.61 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.down_proj_weight_quantizers.62 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.down_proj_weight_quantizers.63 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.64e-02 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.down_proj_weight_quantizers.64 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.67e-01 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.down_proj_weight_quantizers.65 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.28e-01 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.down_proj_weight_quantizers.66 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.down_proj_weight_quantizers.67 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.down_proj_weight_quantizers.68 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.down_proj_weight_quantizers.69 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.down_proj_weight_quantizers.70 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.down_proj_weight_quantizers.71 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.95e-01 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.down_proj_weight_quantizers.72 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.down_proj_weight_quantizers.73 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.down_proj_weight_quantizers.74 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.98e-01 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.down_proj_weight_quantizers.75 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.down_proj_weight_quantizers.76 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.15e-01 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.down_proj_weight_quantizers.77 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.down_proj_weight_quantizers.78 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.down_proj_weight_quantizers.79 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.down_proj_weight_quantizers.80 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.down_proj_weight_quantizers.81 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.down_proj_weight_quantizers.82 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.down_proj_weight_quantizers.83 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.down_proj_weight_quantizers.84 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.down_proj_weight_quantizers.85 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.17e-01 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.down_proj_weight_quantizers.86 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.down_proj_weight_quantizers.87 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.down_proj_weight_quantizers.88 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.down_proj_weight_quantizers.89 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.down_proj_weight_quantizers.90 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.51e-01 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.down_proj_weight_quantizers.91 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.43e-01 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.down_proj_weight_quantizers.92 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.43e-01 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.down_proj_weight_quantizers.93 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.down_proj_weight_quantizers.94 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.down_proj_weight_quantizers.95 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.down_proj_weight_quantizers.96 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.down_proj_weight_quantizers.97 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.51e-01 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.down_proj_weight_quantizers.98 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.35e-01 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.down_proj_weight_quantizers.99 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.down_proj_weight_quantizers.100 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.66e-01 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.down_proj_weight_quantizers.101 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.down_proj_weight_quantizers.102 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.50e-02 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.down_proj_weight_quantizers.103 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.down_proj_weight_quantizers.104 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.down_proj_weight_quantizers.105 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.down_proj_weight_quantizers.106 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.down_proj_weight_quantizers.107 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.40e-02 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.down_proj_weight_quantizers.108 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.19e-01 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.down_proj_weight_quantizers.109 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.down_proj_weight_quantizers.110 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.down_proj_weight_quantizers.111 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.down_proj_weight_quantizers.112 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.down_proj_weight_quantizers.113 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.down_proj_weight_quantizers.114 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.35e-01 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.down_proj_weight_quantizers.115 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.54e-02 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.down_proj_weight_quantizers.116 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.down_proj_weight_quantizers.117 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.59e-01 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.down_proj_weight_quantizers.118 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.19e-01 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.down_proj_weight_quantizers.119 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.down_proj_weight_quantizers.120 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.down_proj_weight_quantizers.121 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.down_proj_weight_quantizers.122 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.down_proj_weight_quantizers.123 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.down_proj_weight_quantizers.124 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.down_proj_weight_quantizers.125 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.down_proj_weight_quantizers.126 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.down_proj_weight_quantizers.127 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.down_proj_weight_quantizers.128 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.45e-02 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.down_proj_weight_quantizers.129 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.down_proj_weight_quantizers.130 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.down_proj_weight_quantizers.131 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.down_proj_weight_quantizers.132 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.down_proj_weight_quantizers.133 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.down_proj_weight_quantizers.134 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.38e-01 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.down_proj_weight_quantizers.135 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.16e-01 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.down_proj_weight_quantizers.136 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.down_proj_weight_quantizers.137 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.down_proj_weight_quantizers.138 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.down_proj_weight_quantizers.139 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.down_proj_weight_quantizers.140 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.down_proj_weight_quantizers.141 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.35e-02 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.down_proj_weight_quantizers.142 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.down_proj_weight_quantizers.143 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.down_proj_weight_quantizers.144 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.down_proj_weight_quantizers.145 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.down_proj_weight_quantizers.146 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.35e-02 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.down_proj_weight_quantizers.147 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.down_proj_weight_quantizers.148 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.down_proj_weight_quantizers.149 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.15e-02 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.down_proj_weight_quantizers.150 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.50e-02 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.down_proj_weight_quantizers.151 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.down_proj_weight_quantizers.152 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.25e-02 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.down_proj_weight_quantizers.153 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.down_proj_weight_quantizers.154 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.down_proj_weight_quantizers.155 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.48e-01 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.down_proj_weight_quantizers.156 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.59e-01 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.down_proj_weight_quantizers.157 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.down_proj_weight_quantizers.158 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.26e-01 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.down_proj_weight_quantizers.159 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.45e-02 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.down_proj_weight_quantizers.160 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.57e-01 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.down_proj_weight_quantizers.161 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.down_proj_weight_quantizers.162 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.down_proj_weight_quantizers.163 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.down_proj_weight_quantizers.164 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.59e-02 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.down_proj_weight_quantizers.165 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.95e-01 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.down_proj_weight_quantizers.166 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.25e-02 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.down_proj_weight_quantizers.167 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.down_proj_weight_quantizers.168 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.64e-01 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.down_proj_weight_quantizers.169 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.19e-01 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.down_proj_weight_quantizers.170 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.down_proj_weight_quantizers.171 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.down_proj_weight_quantizers.172 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.down_proj_weight_quantizers.173 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.54e-02 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.down_proj_weight_quantizers.174 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.11e-01 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.down_proj_weight_quantizers.175 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.down_proj_weight_quantizers.176 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.down_proj_weight_quantizers.177 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.down_proj_weight_quantizers.178 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.down_proj_weight_quantizers.179 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.down_proj_weight_quantizers.180 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.down_proj_weight_quantizers.181 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.down_proj_weight_quantizers.182 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.down_proj_weight_quantizers.183 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.down_proj_weight_quantizers.184 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.29e-01 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.down_proj_weight_quantizers.185 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.down_proj_weight_quantizers.186 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.down_proj_weight_quantizers.187 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.down_proj_weight_quantizers.188 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.03e-01 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.down_proj_weight_quantizers.189 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.95e-01 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.down_proj_weight_quantizers.190 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.down_proj_weight_quantizers.191 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.down_proj_weight_quantizers.192 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.50e-02 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.down_proj_weight_quantizers.193 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.13e-01 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.down_proj_weight_quantizers.194 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.down_proj_weight_quantizers.195 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.down_proj_weight_quantizers.196 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.59e-02 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.down_proj_weight_quantizers.197 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.59e-02 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.down_proj_weight_quantizers.198 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.down_proj_weight_quantizers.199 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.down_proj_weight_quantizers.200 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.down_proj_weight_quantizers.201 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.down_proj_weight_quantizers.202 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.down_proj_weight_quantizers.203 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.down_proj_weight_quantizers.204 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.down_proj_weight_quantizers.205 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.down_proj_weight_quantizers.206 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.71e-01 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.down_proj_weight_quantizers.207 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.down_proj_weight_quantizers.208 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.29e-01 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.down_proj_weight_quantizers.209 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.down_proj_weight_quantizers.210 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.down_proj_weight_quantizers.211 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.29e-01 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.down_proj_weight_quantizers.212 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.down_proj_weight_quantizers.213 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.down_proj_weight_quantizers.214 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.down_proj_weight_quantizers.215 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.down_proj_weight_quantizers.216 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.down_proj_weight_quantizers.217 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.down_proj_weight_quantizers.218 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.down_proj_weight_quantizers.219 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.down_proj_weight_quantizers.220 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.down_proj_weight_quantizers.221 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.29e-01 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.down_proj_weight_quantizers.222 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.down_proj_weight_quantizers.223 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.down_proj_weight_quantizers.224 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.down_proj_weight_quantizers.225 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.down_proj_weight_quantizers.226 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.down_proj_weight_quantizers.227 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.down_proj_weight_quantizers.228 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.down_proj_weight_quantizers.229 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.down_proj_weight_quantizers.230 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.38e-01 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.down_proj_weight_quantizers.231 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.down_proj_weight_quantizers.232 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.45e-02 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.down_proj_weight_quantizers.233 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.down_proj_weight_quantizers.234 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.down_proj_weight_quantizers.235 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.22e-01 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.down_proj_weight_quantizers.236 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.51e-01 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.down_proj_weight_quantizers.237 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.72e-01 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.down_proj_weight_quantizers.238 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.down_proj_weight_quantizers.239 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.17e-01 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.down_proj_weight_quantizers.240 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.down_proj_weight_quantizers.241 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.64e-02 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.down_proj_weight_quantizers.242 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.down_proj_weight_quantizers.243 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.down_proj_weight_quantizers.244 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.down_proj_weight_quantizers.245 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.64e-02 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.down_proj_weight_quantizers.246 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.down_proj_weight_quantizers.247 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.down_proj_weight_quantizers.248 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.down_proj_weight_quantizers.249 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.down_proj_weight_quantizers.250 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.01e-01 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.down_proj_weight_quantizers.251 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.30e-02 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.down_proj_weight_quantizers.252 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.down_proj_weight_quantizers.253 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.down_proj_weight_quantizers.254 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.down_proj_weight_quantizers.255 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.30.mlp.shared_experts.gate_proj.input_quantizer TensorQuantizer(disabled) +model.layers.30.mlp.shared_experts.gate_proj.output_quantizer TensorQuantizer(disabled) +model.layers.30.mlp.shared_experts.gate_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.30.mlp.shared_experts.up_proj.input_quantizer TensorQuantizer(disabled) +model.layers.30.mlp.shared_experts.up_proj.output_quantizer TensorQuantizer(disabled) +model.layers.30.mlp.shared_experts.up_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.30.mlp.shared_experts.down_proj.input_quantizer TensorQuantizer(disabled) +model.layers.30.mlp.shared_experts.down_proj.output_quantizer TensorQuantizer(disabled) +model.layers.30.mlp.shared_experts.down_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.31.self_attn.q_a_proj.input_quantizer TensorQuantizer(disabled) +model.layers.31.self_attn.q_a_proj.output_quantizer TensorQuantizer(disabled) +model.layers.31.self_attn.q_a_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.31.self_attn.q_b_proj.input_quantizer TensorQuantizer(disabled) +model.layers.31.self_attn.q_b_proj.output_quantizer TensorQuantizer(disabled) +model.layers.31.self_attn.q_b_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.31.self_attn.kv_a_proj_with_mqa.input_quantizer TensorQuantizer(disabled) +model.layers.31.self_attn.kv_a_proj_with_mqa.output_quantizer TensorQuantizer(disabled) +model.layers.31.self_attn.kv_a_proj_with_mqa.weight_quantizer TensorQuantizer(disabled) +model.layers.31.self_attn.kv_b_proj.input_quantizer TensorQuantizer(disabled) +model.layers.31.self_attn.kv_b_proj.output_quantizer TensorQuantizer(disabled) +model.layers.31.self_attn.kv_b_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.31.self_attn.o_proj.input_quantizer TensorQuantizer(disabled) +model.layers.31.self_attn.o_proj.output_quantizer TensorQuantizer(disabled) +model.layers.31.self_attn.o_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.31.self_attn.q_bmm_quantizer TensorQuantizer(disabled) +model.layers.31.self_attn.k_bmm_quantizer TensorQuantizer((4, 3) bit fake per-tensor amax=4.48e+02(const) calibrator=MaxCalibrator quant) +model.layers.31.self_attn.v_bmm_quantizer TensorQuantizer((4, 3) bit fake per-tensor amax=4.48e+02(const) calibrator=MaxCalibrator quant) +model.layers.31.self_attn.p_bmm_quantizer TensorQuantizer(disabled) +model.layers.31.mlp.experts.gate_up_proj_input_quantizer TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.30e+00 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.gate_up_proj_weight_quantizers.0 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.57e-01 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.gate_up_proj_weight_quantizers.1 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.91e-01 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.gate_up_proj_weight_quantizers.2 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.60e-01 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.gate_up_proj_weight_quantizers.3 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.47e-01 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.gate_up_proj_weight_quantizers.4 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.gate_up_proj_weight_quantizers.5 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.gate_up_proj_weight_quantizers.6 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.gate_up_proj_weight_quantizers.7 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.gate_up_proj_weight_quantizers.8 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.gate_up_proj_weight_quantizers.9 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.gate_up_proj_weight_quantizers.10 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.gate_up_proj_weight_quantizers.11 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.14e-01 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.gate_up_proj_weight_quantizers.12 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.gate_up_proj_weight_quantizers.13 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.gate_up_proj_weight_quantizers.14 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.gate_up_proj_weight_quantizers.15 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.03e-01 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.gate_up_proj_weight_quantizers.16 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.47e-01 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.gate_up_proj_weight_quantizers.17 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.gate_up_proj_weight_quantizers.18 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.gate_up_proj_weight_quantizers.19 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.13e-01 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.gate_up_proj_weight_quantizers.20 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.54e-01 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.gate_up_proj_weight_quantizers.21 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.gate_up_proj_weight_quantizers.22 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.gate_up_proj_weight_quantizers.23 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.gate_up_proj_weight_quantizers.24 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.gate_up_proj_weight_quantizers.25 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.gate_up_proj_weight_quantizers.26 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.16e-01 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.gate_up_proj_weight_quantizers.27 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.31e-01 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.gate_up_proj_weight_quantizers.28 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.gate_up_proj_weight_quantizers.29 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.30e-01 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.gate_up_proj_weight_quantizers.30 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.22e-01 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.gate_up_proj_weight_quantizers.31 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.gate_up_proj_weight_quantizers.32 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.25e-01 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.gate_up_proj_weight_quantizers.33 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.gate_up_proj_weight_quantizers.34 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.54e-01 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.gate_up_proj_weight_quantizers.35 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.gate_up_proj_weight_quantizers.36 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.gate_up_proj_weight_quantizers.37 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.gate_up_proj_weight_quantizers.38 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.gate_up_proj_weight_quantizers.39 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.gate_up_proj_weight_quantizers.40 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.47e-01 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.gate_up_proj_weight_quantizers.41 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.gate_up_proj_weight_quantizers.42 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.gate_up_proj_weight_quantizers.43 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.gate_up_proj_weight_quantizers.44 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.gate_up_proj_weight_quantizers.45 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.15e-01 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.gate_up_proj_weight_quantizers.46 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.gate_up_proj_weight_quantizers.47 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.gate_up_proj_weight_quantizers.48 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.gate_up_proj_weight_quantizers.49 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.33e-01 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.gate_up_proj_weight_quantizers.50 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.gate_up_proj_weight_quantizers.51 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.54e-02 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.gate_up_proj_weight_quantizers.52 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.gate_up_proj_weight_quantizers.53 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.gate_up_proj_weight_quantizers.54 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.gate_up_proj_weight_quantizers.55 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.gate_up_proj_weight_quantizers.56 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.gate_up_proj_weight_quantizers.57 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.gate_up_proj_weight_quantizers.58 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.gate_up_proj_weight_quantizers.59 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.gate_up_proj_weight_quantizers.60 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.gate_up_proj_weight_quantizers.61 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.gate_up_proj_weight_quantizers.62 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.15e-01 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.gate_up_proj_weight_quantizers.63 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.51e-01 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.gate_up_proj_weight_quantizers.64 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.49e-01 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.gate_up_proj_weight_quantizers.65 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.68e-01 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.gate_up_proj_weight_quantizers.66 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.gate_up_proj_weight_quantizers.67 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.18e-01 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.gate_up_proj_weight_quantizers.68 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.53e-01 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.gate_up_proj_weight_quantizers.69 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.25e-01 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.gate_up_proj_weight_quantizers.70 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.gate_up_proj_weight_quantizers.71 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.gate_up_proj_weight_quantizers.72 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.gate_up_proj_weight_quantizers.73 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.gate_up_proj_weight_quantizers.74 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.gate_up_proj_weight_quantizers.75 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.20e-01 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.gate_up_proj_weight_quantizers.76 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.00e-01 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.gate_up_proj_weight_quantizers.77 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.gate_up_proj_weight_quantizers.78 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.gate_up_proj_weight_quantizers.79 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.gate_up_proj_weight_quantizers.80 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.37e-01 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.gate_up_proj_weight_quantizers.81 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.gate_up_proj_weight_quantizers.82 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.gate_up_proj_weight_quantizers.83 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.gate_up_proj_weight_quantizers.84 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.gate_up_proj_weight_quantizers.85 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.gate_up_proj_weight_quantizers.86 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.29e-01 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.gate_up_proj_weight_quantizers.87 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.gate_up_proj_weight_quantizers.88 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.gate_up_proj_weight_quantizers.89 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.40e-01 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.gate_up_proj_weight_quantizers.90 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.46e-01 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.gate_up_proj_weight_quantizers.91 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.gate_up_proj_weight_quantizers.92 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.43e-01 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.gate_up_proj_weight_quantizers.93 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.gate_up_proj_weight_quantizers.94 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.gate_up_proj_weight_quantizers.95 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.18e-01 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.gate_up_proj_weight_quantizers.96 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.68e-01 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.gate_up_proj_weight_quantizers.97 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.53e-01 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.gate_up_proj_weight_quantizers.98 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.gate_up_proj_weight_quantizers.99 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.gate_up_proj_weight_quantizers.100 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.gate_up_proj_weight_quantizers.101 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.gate_up_proj_weight_quantizers.102 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.57e-01 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.gate_up_proj_weight_quantizers.103 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.gate_up_proj_weight_quantizers.104 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.18e-01 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.gate_up_proj_weight_quantizers.105 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.13e-01 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.gate_up_proj_weight_quantizers.106 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.gate_up_proj_weight_quantizers.107 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.gate_up_proj_weight_quantizers.108 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.13e-01 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.gate_up_proj_weight_quantizers.109 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.gate_up_proj_weight_quantizers.110 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.16e-01 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.gate_up_proj_weight_quantizers.111 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.gate_up_proj_weight_quantizers.112 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.gate_up_proj_weight_quantizers.113 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.gate_up_proj_weight_quantizers.114 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.gate_up_proj_weight_quantizers.115 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.gate_up_proj_weight_quantizers.116 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.gate_up_proj_weight_quantizers.117 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.gate_up_proj_weight_quantizers.118 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.60e-01 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.gate_up_proj_weight_quantizers.119 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.gate_up_proj_weight_quantizers.120 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.gate_up_proj_weight_quantizers.121 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.gate_up_proj_weight_quantizers.122 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.gate_up_proj_weight_quantizers.123 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.gate_up_proj_weight_quantizers.124 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.gate_up_proj_weight_quantizers.125 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.gate_up_proj_weight_quantizers.126 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.20e-01 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.gate_up_proj_weight_quantizers.127 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.gate_up_proj_weight_quantizers.128 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.gate_up_proj_weight_quantizers.129 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.gate_up_proj_weight_quantizers.130 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.22e-01 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.gate_up_proj_weight_quantizers.131 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.33e-01 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.gate_up_proj_weight_quantizers.132 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.gate_up_proj_weight_quantizers.133 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.gate_up_proj_weight_quantizers.134 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.gate_up_proj_weight_quantizers.135 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.43e-01 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.gate_up_proj_weight_quantizers.136 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.46e-01 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.gate_up_proj_weight_quantizers.137 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.gate_up_proj_weight_quantizers.138 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.gate_up_proj_weight_quantizers.139 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.gate_up_proj_weight_quantizers.140 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.gate_up_proj_weight_quantizers.141 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.gate_up_proj_weight_quantizers.142 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.gate_up_proj_weight_quantizers.143 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.40e-01 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.gate_up_proj_weight_quantizers.144 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.gate_up_proj_weight_quantizers.145 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.gate_up_proj_weight_quantizers.146 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.gate_up_proj_weight_quantizers.147 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.16e-01 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.gate_up_proj_weight_quantizers.148 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.gate_up_proj_weight_quantizers.149 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.gate_up_proj_weight_quantizers.150 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.gate_up_proj_weight_quantizers.151 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.43e-01 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.gate_up_proj_weight_quantizers.152 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.gate_up_proj_weight_quantizers.153 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.41e-01 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.gate_up_proj_weight_quantizers.154 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.gate_up_proj_weight_quantizers.155 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.gate_up_proj_weight_quantizers.156 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.gate_up_proj_weight_quantizers.157 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.gate_up_proj_weight_quantizers.158 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.24e-01 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.gate_up_proj_weight_quantizers.159 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.gate_up_proj_weight_quantizers.160 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.14e-01 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.gate_up_proj_weight_quantizers.161 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.gate_up_proj_weight_quantizers.162 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.gate_up_proj_weight_quantizers.163 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.gate_up_proj_weight_quantizers.164 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.gate_up_proj_weight_quantizers.165 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.gate_up_proj_weight_quantizers.166 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.49e-01 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.gate_up_proj_weight_quantizers.167 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.gate_up_proj_weight_quantizers.168 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.66e-01 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.gate_up_proj_weight_quantizers.169 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.gate_up_proj_weight_quantizers.170 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.87e-01 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.gate_up_proj_weight_quantizers.171 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.gate_up_proj_weight_quantizers.172 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.gate_up_proj_weight_quantizers.173 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.gate_up_proj_weight_quantizers.174 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.gate_up_proj_weight_quantizers.175 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.gate_up_proj_weight_quantizers.176 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.gate_up_proj_weight_quantizers.177 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.13e-01 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.gate_up_proj_weight_quantizers.178 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.88e-01 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.gate_up_proj_weight_quantizers.179 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.gate_up_proj_weight_quantizers.180 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.gate_up_proj_weight_quantizers.181 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.gate_up_proj_weight_quantizers.182 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.gate_up_proj_weight_quantizers.183 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.gate_up_proj_weight_quantizers.184 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.87e-01 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.gate_up_proj_weight_quantizers.185 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.15e-01 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.gate_up_proj_weight_quantizers.186 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.gate_up_proj_weight_quantizers.187 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.33e-01 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.gate_up_proj_weight_quantizers.188 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.gate_up_proj_weight_quantizers.189 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.gate_up_proj_weight_quantizers.190 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.gate_up_proj_weight_quantizers.191 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.gate_up_proj_weight_quantizers.192 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.65e-01 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.gate_up_proj_weight_quantizers.193 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.11e-01 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.gate_up_proj_weight_quantizers.194 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.21e-01 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.gate_up_proj_weight_quantizers.195 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.gate_up_proj_weight_quantizers.196 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.gate_up_proj_weight_quantizers.197 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.gate_up_proj_weight_quantizers.198 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.22e-01 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.gate_up_proj_weight_quantizers.199 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.38e-01 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.gate_up_proj_weight_quantizers.200 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.gate_up_proj_weight_quantizers.201 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.gate_up_proj_weight_quantizers.202 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.gate_up_proj_weight_quantizers.203 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.gate_up_proj_weight_quantizers.204 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.24e-01 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.gate_up_proj_weight_quantizers.205 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.gate_up_proj_weight_quantizers.206 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.gate_up_proj_weight_quantizers.207 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.gate_up_proj_weight_quantizers.208 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.gate_up_proj_weight_quantizers.209 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.16e-01 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.gate_up_proj_weight_quantizers.210 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.50e-01 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.gate_up_proj_weight_quantizers.211 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.46e-01 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.gate_up_proj_weight_quantizers.212 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.gate_up_proj_weight_quantizers.213 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.13e-01 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.gate_up_proj_weight_quantizers.214 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.19e-01 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.gate_up_proj_weight_quantizers.215 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.59e-02 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.gate_up_proj_weight_quantizers.216 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.gate_up_proj_weight_quantizers.217 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.gate_up_proj_weight_quantizers.218 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.gate_up_proj_weight_quantizers.219 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.64e-02 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.gate_up_proj_weight_quantizers.220 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.51e-01 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.gate_up_proj_weight_quantizers.221 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.20e-01 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.gate_up_proj_weight_quantizers.222 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.gate_up_proj_weight_quantizers.223 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.98e-01 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.gate_up_proj_weight_quantizers.224 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.gate_up_proj_weight_quantizers.225 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.gate_up_proj_weight_quantizers.226 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.gate_up_proj_weight_quantizers.227 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.gate_up_proj_weight_quantizers.228 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.gate_up_proj_weight_quantizers.229 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.gate_up_proj_weight_quantizers.230 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.34e-01 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.gate_up_proj_weight_quantizers.231 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.gate_up_proj_weight_quantizers.232 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.gate_up_proj_weight_quantizers.233 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.gate_up_proj_weight_quantizers.234 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.38e-01 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.gate_up_proj_weight_quantizers.235 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.26e-01 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.gate_up_proj_weight_quantizers.236 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.gate_up_proj_weight_quantizers.237 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.74e-01 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.gate_up_proj_weight_quantizers.238 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.32e-01 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.gate_up_proj_weight_quantizers.239 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.gate_up_proj_weight_quantizers.240 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.gate_up_proj_weight_quantizers.241 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.gate_up_proj_weight_quantizers.242 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.08e-01 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.gate_up_proj_weight_quantizers.243 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.gate_up_proj_weight_quantizers.244 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.gate_up_proj_weight_quantizers.245 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.38e-01 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.gate_up_proj_weight_quantizers.246 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.gate_up_proj_weight_quantizers.247 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.gate_up_proj_weight_quantizers.248 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.49e-01 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.gate_up_proj_weight_quantizers.249 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.gate_up_proj_weight_quantizers.250 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.64e-01 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.gate_up_proj_weight_quantizers.251 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.13e-01 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.gate_up_proj_weight_quantizers.252 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.gate_up_proj_weight_quantizers.253 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.gate_up_proj_weight_quantizers.254 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.gate_up_proj_weight_quantizers.255 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.down_proj_input_quantizer TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.42e+01 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.down_proj_weight_quantizers.0 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.down_proj_weight_quantizers.1 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.down_proj_weight_quantizers.2 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.down_proj_weight_quantizers.3 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.18e-01 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.down_proj_weight_quantizers.4 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.down_proj_weight_quantizers.5 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.down_proj_weight_quantizers.6 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.down_proj_weight_quantizers.7 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.down_proj_weight_quantizers.8 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.down_proj_weight_quantizers.9 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.down_proj_weight_quantizers.10 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.down_proj_weight_quantizers.11 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.down_proj_weight_quantizers.12 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.down_proj_weight_quantizers.13 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.down_proj_weight_quantizers.14 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.down_proj_weight_quantizers.15 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.down_proj_weight_quantizers.16 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.down_proj_weight_quantizers.17 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.down_proj_weight_quantizers.18 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.64e-02 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.down_proj_weight_quantizers.19 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.down_proj_weight_quantizers.20 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.29e-01 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.down_proj_weight_quantizers.21 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.64e-02 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.down_proj_weight_quantizers.22 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.down_proj_weight_quantizers.23 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.down_proj_weight_quantizers.24 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.down_proj_weight_quantizers.25 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.down_proj_weight_quantizers.26 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.down_proj_weight_quantizers.27 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.24e-01 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.down_proj_weight_quantizers.28 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.down_proj_weight_quantizers.29 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.down_proj_weight_quantizers.30 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.down_proj_weight_quantizers.31 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.down_proj_weight_quantizers.32 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.19e-01 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.down_proj_weight_quantizers.33 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.down_proj_weight_quantizers.34 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.87e-01 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.down_proj_weight_quantizers.35 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.down_proj_weight_quantizers.36 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.59e-02 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.down_proj_weight_quantizers.37 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.down_proj_weight_quantizers.38 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.30e-02 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.down_proj_weight_quantizers.39 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.down_proj_weight_quantizers.40 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.down_proj_weight_quantizers.41 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.down_proj_weight_quantizers.42 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.down_proj_weight_quantizers.43 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.down_proj_weight_quantizers.44 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.down_proj_weight_quantizers.45 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.down_proj_weight_quantizers.46 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.down_proj_weight_quantizers.47 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.down_proj_weight_quantizers.48 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.down_proj_weight_quantizers.49 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.down_proj_weight_quantizers.50 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.down_proj_weight_quantizers.51 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.down_proj_weight_quantizers.52 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.down_proj_weight_quantizers.53 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.down_proj_weight_quantizers.54 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.down_proj_weight_quantizers.55 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.down_proj_weight_quantizers.56 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.down_proj_weight_quantizers.57 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.down_proj_weight_quantizers.58 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.down_proj_weight_quantizers.59 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.down_proj_weight_quantizers.60 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.down_proj_weight_quantizers.61 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.down_proj_weight_quantizers.62 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.25e-02 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.down_proj_weight_quantizers.63 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.down_proj_weight_quantizers.64 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.25e-01 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.down_proj_weight_quantizers.65 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.46e-01 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.down_proj_weight_quantizers.66 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.30e-02 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.down_proj_weight_quantizers.67 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.59e-02 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.down_proj_weight_quantizers.68 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.down_proj_weight_quantizers.69 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.22e-01 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.down_proj_weight_quantizers.70 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.down_proj_weight_quantizers.71 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.59e-02 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.down_proj_weight_quantizers.72 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.down_proj_weight_quantizers.73 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.down_proj_weight_quantizers.74 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.down_proj_weight_quantizers.75 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.14e-01 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.down_proj_weight_quantizers.76 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.down_proj_weight_quantizers.77 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.down_proj_weight_quantizers.78 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.down_proj_weight_quantizers.79 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.down_proj_weight_quantizers.80 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.down_proj_weight_quantizers.81 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.down_proj_weight_quantizers.82 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.down_proj_weight_quantizers.83 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.down_proj_weight_quantizers.84 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.40e-02 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.down_proj_weight_quantizers.85 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.down_proj_weight_quantizers.86 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.54e-02 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.down_proj_weight_quantizers.87 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.down_proj_weight_quantizers.88 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.down_proj_weight_quantizers.89 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.64e-02 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.down_proj_weight_quantizers.90 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.75e-01 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.down_proj_weight_quantizers.91 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.down_proj_weight_quantizers.92 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.17e-01 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.down_proj_weight_quantizers.93 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.54e-02 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.down_proj_weight_quantizers.94 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.20e-02 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.down_proj_weight_quantizers.95 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.down_proj_weight_quantizers.96 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.down_proj_weight_quantizers.97 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.61e-01 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.down_proj_weight_quantizers.98 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.down_proj_weight_quantizers.99 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.down_proj_weight_quantizers.100 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.down_proj_weight_quantizers.101 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.39e-01 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.down_proj_weight_quantizers.102 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.88e-01 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.down_proj_weight_quantizers.103 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.down_proj_weight_quantizers.104 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.down_proj_weight_quantizers.105 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.down_proj_weight_quantizers.106 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.down_proj_weight_quantizers.107 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.down_proj_weight_quantizers.108 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.down_proj_weight_quantizers.109 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.35e-02 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.down_proj_weight_quantizers.110 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.down_proj_weight_quantizers.111 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.64e-02 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.down_proj_weight_quantizers.112 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.down_proj_weight_quantizers.113 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.down_proj_weight_quantizers.114 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.down_proj_weight_quantizers.115 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.27e-01 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.down_proj_weight_quantizers.116 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.down_proj_weight_quantizers.117 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.down_proj_weight_quantizers.118 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.52e-01 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.down_proj_weight_quantizers.119 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.34e-01 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.down_proj_weight_quantizers.120 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.down_proj_weight_quantizers.121 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.down_proj_weight_quantizers.122 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.down_proj_weight_quantizers.123 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.down_proj_weight_quantizers.124 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.down_proj_weight_quantizers.125 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.down_proj_weight_quantizers.126 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.down_proj_weight_quantizers.127 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.down_proj_weight_quantizers.128 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.down_proj_weight_quantizers.129 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.down_proj_weight_quantizers.130 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.87e-01 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.down_proj_weight_quantizers.131 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.down_proj_weight_quantizers.132 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.18e-01 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.down_proj_weight_quantizers.133 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.down_proj_weight_quantizers.134 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.down_proj_weight_quantizers.135 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.down_proj_weight_quantizers.136 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.down_proj_weight_quantizers.137 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.16e-01 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.down_proj_weight_quantizers.138 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.59e-02 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.down_proj_weight_quantizers.139 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.down_proj_weight_quantizers.140 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.down_proj_weight_quantizers.141 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.down_proj_weight_quantizers.142 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.down_proj_weight_quantizers.143 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.17e-01 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.down_proj_weight_quantizers.144 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.down_proj_weight_quantizers.145 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.40e-02 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.down_proj_weight_quantizers.146 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.59e-02 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.down_proj_weight_quantizers.147 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.down_proj_weight_quantizers.148 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.down_proj_weight_quantizers.149 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.64e-02 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.down_proj_weight_quantizers.150 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.50e-02 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.down_proj_weight_quantizers.151 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.down_proj_weight_quantizers.152 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.down_proj_weight_quantizers.153 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.down_proj_weight_quantizers.154 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.50e-02 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.down_proj_weight_quantizers.155 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.down_proj_weight_quantizers.156 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.45e-02 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.down_proj_weight_quantizers.157 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.down_proj_weight_quantizers.158 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.77e-01 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.down_proj_weight_quantizers.159 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.64e-02 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.down_proj_weight_quantizers.160 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.50e-02 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.down_proj_weight_quantizers.161 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.31e-01 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.down_proj_weight_quantizers.162 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.down_proj_weight_quantizers.163 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.down_proj_weight_quantizers.164 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.down_proj_weight_quantizers.165 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.down_proj_weight_quantizers.166 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.41e-01 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.down_proj_weight_quantizers.167 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.down_proj_weight_quantizers.168 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.down_proj_weight_quantizers.169 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.down_proj_weight_quantizers.170 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.54e-01 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.down_proj_weight_quantizers.171 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.down_proj_weight_quantizers.172 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.54e-02 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.down_proj_weight_quantizers.173 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.down_proj_weight_quantizers.174 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.63e-01 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.down_proj_weight_quantizers.175 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.64e-02 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.down_proj_weight_quantizers.176 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.down_proj_weight_quantizers.177 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.down_proj_weight_quantizers.178 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.down_proj_weight_quantizers.179 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.down_proj_weight_quantizers.180 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.down_proj_weight_quantizers.181 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.40e-02 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.down_proj_weight_quantizers.182 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.down_proj_weight_quantizers.183 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.down_proj_weight_quantizers.184 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.26e-01 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.down_proj_weight_quantizers.185 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.down_proj_weight_quantizers.186 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.down_proj_weight_quantizers.187 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.down_proj_weight_quantizers.188 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.down_proj_weight_quantizers.189 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.down_proj_weight_quantizers.190 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.down_proj_weight_quantizers.191 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.64e-02 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.down_proj_weight_quantizers.192 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=5.16e-01 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.down_proj_weight_quantizers.193 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.57e-01 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.down_proj_weight_quantizers.194 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.down_proj_weight_quantizers.195 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.down_proj_weight_quantizers.196 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.down_proj_weight_quantizers.197 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.down_proj_weight_quantizers.198 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.down_proj_weight_quantizers.199 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.down_proj_weight_quantizers.200 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.down_proj_weight_quantizers.201 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.64e-02 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.down_proj_weight_quantizers.202 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.down_proj_weight_quantizers.203 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.down_proj_weight_quantizers.204 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.down_proj_weight_quantizers.205 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.down_proj_weight_quantizers.206 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.down_proj_weight_quantizers.207 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.down_proj_weight_quantizers.208 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.59e-02 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.down_proj_weight_quantizers.209 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.down_proj_weight_quantizers.210 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.45e-01 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.down_proj_weight_quantizers.211 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.down_proj_weight_quantizers.212 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.down_proj_weight_quantizers.213 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.down_proj_weight_quantizers.214 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.down_proj_weight_quantizers.215 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.64e-02 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.down_proj_weight_quantizers.216 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.down_proj_weight_quantizers.217 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.64e-02 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.down_proj_weight_quantizers.218 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.down_proj_weight_quantizers.219 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.59e-02 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.down_proj_weight_quantizers.220 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.35e-01 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.down_proj_weight_quantizers.221 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.down_proj_weight_quantizers.222 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.down_proj_weight_quantizers.223 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.25e-01 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.down_proj_weight_quantizers.224 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.down_proj_weight_quantizers.225 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.down_proj_weight_quantizers.226 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.down_proj_weight_quantizers.227 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.down_proj_weight_quantizers.228 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.down_proj_weight_quantizers.229 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.40e-02 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.down_proj_weight_quantizers.230 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.down_proj_weight_quantizers.231 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.35e-02 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.down_proj_weight_quantizers.232 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.down_proj_weight_quantizers.233 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.down_proj_weight_quantizers.234 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.down_proj_weight_quantizers.235 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.down_proj_weight_quantizers.236 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.down_proj_weight_quantizers.237 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.down_proj_weight_quantizers.238 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.63e-01 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.down_proj_weight_quantizers.239 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.down_proj_weight_quantizers.240 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.down_proj_weight_quantizers.241 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.down_proj_weight_quantizers.242 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.down_proj_weight_quantizers.243 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.down_proj_weight_quantizers.244 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.down_proj_weight_quantizers.245 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.down_proj_weight_quantizers.246 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.35e-02 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.down_proj_weight_quantizers.247 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.59e-02 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.down_proj_weight_quantizers.248 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.20e-01 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.down_proj_weight_quantizers.249 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.54e-02 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.down_proj_weight_quantizers.250 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.down_proj_weight_quantizers.251 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.59e-02 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.down_proj_weight_quantizers.252 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.down_proj_weight_quantizers.253 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.45e-02 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.down_proj_weight_quantizers.254 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.down_proj_weight_quantizers.255 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.31.mlp.shared_experts.gate_proj.input_quantizer TensorQuantizer(disabled) +model.layers.31.mlp.shared_experts.gate_proj.output_quantizer TensorQuantizer(disabled) +model.layers.31.mlp.shared_experts.gate_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.31.mlp.shared_experts.up_proj.input_quantizer TensorQuantizer(disabled) +model.layers.31.mlp.shared_experts.up_proj.output_quantizer TensorQuantizer(disabled) +model.layers.31.mlp.shared_experts.up_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.31.mlp.shared_experts.down_proj.input_quantizer TensorQuantizer(disabled) +model.layers.31.mlp.shared_experts.down_proj.output_quantizer TensorQuantizer(disabled) +model.layers.31.mlp.shared_experts.down_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.32.self_attn.q_a_proj.input_quantizer TensorQuantizer(disabled) +model.layers.32.self_attn.q_a_proj.output_quantizer TensorQuantizer(disabled) +model.layers.32.self_attn.q_a_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.32.self_attn.q_b_proj.input_quantizer TensorQuantizer(disabled) +model.layers.32.self_attn.q_b_proj.output_quantizer TensorQuantizer(disabled) +model.layers.32.self_attn.q_b_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.32.self_attn.kv_a_proj_with_mqa.input_quantizer TensorQuantizer(disabled) +model.layers.32.self_attn.kv_a_proj_with_mqa.output_quantizer TensorQuantizer(disabled) +model.layers.32.self_attn.kv_a_proj_with_mqa.weight_quantizer TensorQuantizer(disabled) +model.layers.32.self_attn.kv_b_proj.input_quantizer TensorQuantizer(disabled) +model.layers.32.self_attn.kv_b_proj.output_quantizer TensorQuantizer(disabled) +model.layers.32.self_attn.kv_b_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.32.self_attn.o_proj.input_quantizer TensorQuantizer(disabled) +model.layers.32.self_attn.o_proj.output_quantizer TensorQuantizer(disabled) +model.layers.32.self_attn.o_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.32.self_attn.q_bmm_quantizer TensorQuantizer(disabled) +model.layers.32.self_attn.k_bmm_quantizer TensorQuantizer((4, 3) bit fake per-tensor amax=4.48e+02(const) calibrator=MaxCalibrator quant) +model.layers.32.self_attn.v_bmm_quantizer TensorQuantizer((4, 3) bit fake per-tensor amax=4.48e+02(const) calibrator=MaxCalibrator quant) +model.layers.32.self_attn.p_bmm_quantizer TensorQuantizer(disabled) +model.layers.32.mlp.experts.gate_up_proj_input_quantizer TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.27e+00 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.gate_up_proj_weight_quantizers.0 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.42e-01 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.gate_up_proj_weight_quantizers.1 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.gate_up_proj_weight_quantizers.2 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.gate_up_proj_weight_quantizers.3 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.gate_up_proj_weight_quantizers.4 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.gate_up_proj_weight_quantizers.5 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.gate_up_proj_weight_quantizers.6 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.gate_up_proj_weight_quantizers.7 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.gate_up_proj_weight_quantizers.8 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.85e-01 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.gate_up_proj_weight_quantizers.9 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.gate_up_proj_weight_quantizers.10 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.50e-01 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.gate_up_proj_weight_quantizers.11 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.13e-01 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.gate_up_proj_weight_quantizers.12 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.gate_up_proj_weight_quantizers.13 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.21e-01 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.gate_up_proj_weight_quantizers.14 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.31e-01 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.gate_up_proj_weight_quantizers.15 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.gate_up_proj_weight_quantizers.16 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.45e-01 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.gate_up_proj_weight_quantizers.17 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.gate_up_proj_weight_quantizers.18 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.48e-01 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.gate_up_proj_weight_quantizers.19 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.gate_up_proj_weight_quantizers.20 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.58e-01 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.gate_up_proj_weight_quantizers.21 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.gate_up_proj_weight_quantizers.22 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.gate_up_proj_weight_quantizers.23 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.gate_up_proj_weight_quantizers.24 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.gate_up_proj_weight_quantizers.25 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.gate_up_proj_weight_quantizers.26 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.20e-01 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.gate_up_proj_weight_quantizers.27 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.41e-01 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.gate_up_proj_weight_quantizers.28 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.gate_up_proj_weight_quantizers.29 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.gate_up_proj_weight_quantizers.30 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.gate_up_proj_weight_quantizers.31 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.gate_up_proj_weight_quantizers.32 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.gate_up_proj_weight_quantizers.33 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.gate_up_proj_weight_quantizers.34 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.gate_up_proj_weight_quantizers.35 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.gate_up_proj_weight_quantizers.36 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.gate_up_proj_weight_quantizers.37 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.gate_up_proj_weight_quantizers.38 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.gate_up_proj_weight_quantizers.39 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.gate_up_proj_weight_quantizers.40 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.gate_up_proj_weight_quantizers.41 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.02e-01 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.gate_up_proj_weight_quantizers.42 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.14e-01 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.gate_up_proj_weight_quantizers.43 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.gate_up_proj_weight_quantizers.44 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.90e-01 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.gate_up_proj_weight_quantizers.45 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.gate_up_proj_weight_quantizers.46 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.gate_up_proj_weight_quantizers.47 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.gate_up_proj_weight_quantizers.48 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.gate_up_proj_weight_quantizers.49 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.gate_up_proj_weight_quantizers.50 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.gate_up_proj_weight_quantizers.51 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.gate_up_proj_weight_quantizers.52 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.gate_up_proj_weight_quantizers.53 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.67e-01 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.gate_up_proj_weight_quantizers.54 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.gate_up_proj_weight_quantizers.55 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.gate_up_proj_weight_quantizers.56 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.40e-01 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.gate_up_proj_weight_quantizers.57 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.gate_up_proj_weight_quantizers.58 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.gate_up_proj_weight_quantizers.59 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.gate_up_proj_weight_quantizers.60 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.80e-01 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.gate_up_proj_weight_quantizers.61 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.gate_up_proj_weight_quantizers.62 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.46e-01 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.gate_up_proj_weight_quantizers.63 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.gate_up_proj_weight_quantizers.64 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.gate_up_proj_weight_quantizers.65 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.45e-01 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.gate_up_proj_weight_quantizers.66 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.73e-01 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.gate_up_proj_weight_quantizers.67 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.gate_up_proj_weight_quantizers.68 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.gate_up_proj_weight_quantizers.69 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.07e-01 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.gate_up_proj_weight_quantizers.70 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.gate_up_proj_weight_quantizers.71 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.22e-01 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.gate_up_proj_weight_quantizers.72 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.33e-01 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.gate_up_proj_weight_quantizers.73 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.gate_up_proj_weight_quantizers.74 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.46e-01 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.gate_up_proj_weight_quantizers.75 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.gate_up_proj_weight_quantizers.76 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.29e-01 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.gate_up_proj_weight_quantizers.77 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.24e-01 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.gate_up_proj_weight_quantizers.78 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.gate_up_proj_weight_quantizers.79 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.gate_up_proj_weight_quantizers.80 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.gate_up_proj_weight_quantizers.81 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.25e-01 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.gate_up_proj_weight_quantizers.82 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.gate_up_proj_weight_quantizers.83 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.81e-01 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.gate_up_proj_weight_quantizers.84 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.30e-01 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.gate_up_proj_weight_quantizers.85 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.gate_up_proj_weight_quantizers.86 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.gate_up_proj_weight_quantizers.87 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.gate_up_proj_weight_quantizers.88 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.gate_up_proj_weight_quantizers.89 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.31e-01 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.gate_up_proj_weight_quantizers.90 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.gate_up_proj_weight_quantizers.91 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.gate_up_proj_weight_quantizers.92 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.05e-01 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.gate_up_proj_weight_quantizers.93 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.gate_up_proj_weight_quantizers.94 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.24e-01 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.gate_up_proj_weight_quantizers.95 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.42e-01 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.gate_up_proj_weight_quantizers.96 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.gate_up_proj_weight_quantizers.97 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.gate_up_proj_weight_quantizers.98 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.03e-01 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.gate_up_proj_weight_quantizers.99 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.35e-01 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.gate_up_proj_weight_quantizers.100 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.19e-01 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.gate_up_proj_weight_quantizers.101 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.88e-01 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.gate_up_proj_weight_quantizers.102 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.gate_up_proj_weight_quantizers.103 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.gate_up_proj_weight_quantizers.104 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.gate_up_proj_weight_quantizers.105 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.gate_up_proj_weight_quantizers.106 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.26e-01 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.gate_up_proj_weight_quantizers.107 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.13e-01 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.gate_up_proj_weight_quantizers.108 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.gate_up_proj_weight_quantizers.109 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.gate_up_proj_weight_quantizers.110 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.04e-01 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.gate_up_proj_weight_quantizers.111 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.14e-01 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.gate_up_proj_weight_quantizers.112 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.22e-01 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.gate_up_proj_weight_quantizers.113 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.75e-01 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.gate_up_proj_weight_quantizers.114 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.18e-01 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.gate_up_proj_weight_quantizers.115 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.19e-01 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.gate_up_proj_weight_quantizers.116 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.34e-01 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.gate_up_proj_weight_quantizers.117 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.27e-01 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.gate_up_proj_weight_quantizers.118 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.25e-01 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.gate_up_proj_weight_quantizers.119 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.gate_up_proj_weight_quantizers.120 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.gate_up_proj_weight_quantizers.121 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.gate_up_proj_weight_quantizers.122 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.gate_up_proj_weight_quantizers.123 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.gate_up_proj_weight_quantizers.124 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.36e-01 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.gate_up_proj_weight_quantizers.125 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.gate_up_proj_weight_quantizers.126 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.18e-01 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.gate_up_proj_weight_quantizers.127 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.20e-01 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.gate_up_proj_weight_quantizers.128 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.gate_up_proj_weight_quantizers.129 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.15e-01 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.gate_up_proj_weight_quantizers.130 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.gate_up_proj_weight_quantizers.131 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.36e-01 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.gate_up_proj_weight_quantizers.132 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.gate_up_proj_weight_quantizers.133 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.32e-01 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.gate_up_proj_weight_quantizers.134 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.gate_up_proj_weight_quantizers.135 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.gate_up_proj_weight_quantizers.136 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.gate_up_proj_weight_quantizers.137 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.24e-01 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.gate_up_proj_weight_quantizers.138 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.57e-01 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.gate_up_proj_weight_quantizers.139 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.25e-01 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.gate_up_proj_weight_quantizers.140 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.gate_up_proj_weight_quantizers.141 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.19e-01 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.gate_up_proj_weight_quantizers.142 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.gate_up_proj_weight_quantizers.143 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.40e-01 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.gate_up_proj_weight_quantizers.144 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.75e-01 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.gate_up_proj_weight_quantizers.145 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.63e-01 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.gate_up_proj_weight_quantizers.146 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.31e-01 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.gate_up_proj_weight_quantizers.147 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.15e-01 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.gate_up_proj_weight_quantizers.148 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.gate_up_proj_weight_quantizers.149 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.24e-01 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.gate_up_proj_weight_quantizers.150 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.gate_up_proj_weight_quantizers.151 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.17e-01 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.gate_up_proj_weight_quantizers.152 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.gate_up_proj_weight_quantizers.153 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.15e-01 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.gate_up_proj_weight_quantizers.154 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.gate_up_proj_weight_quantizers.155 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.gate_up_proj_weight_quantizers.156 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.gate_up_proj_weight_quantizers.157 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.gate_up_proj_weight_quantizers.158 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.gate_up_proj_weight_quantizers.159 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.gate_up_proj_weight_quantizers.160 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.gate_up_proj_weight_quantizers.161 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.gate_up_proj_weight_quantizers.162 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.gate_up_proj_weight_quantizers.163 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.gate_up_proj_weight_quantizers.164 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.17e-01 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.gate_up_proj_weight_quantizers.165 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.gate_up_proj_weight_quantizers.166 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.gate_up_proj_weight_quantizers.167 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.09e-01 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.gate_up_proj_weight_quantizers.168 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.gate_up_proj_weight_quantizers.169 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.gate_up_proj_weight_quantizers.170 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.gate_up_proj_weight_quantizers.171 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.gate_up_proj_weight_quantizers.172 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.gate_up_proj_weight_quantizers.173 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.20e-01 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.gate_up_proj_weight_quantizers.174 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.gate_up_proj_weight_quantizers.175 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.66e-01 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.gate_up_proj_weight_quantizers.176 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.gate_up_proj_weight_quantizers.177 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.25e-01 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.gate_up_proj_weight_quantizers.178 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.21e-01 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.gate_up_proj_weight_quantizers.179 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.19e-01 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.gate_up_proj_weight_quantizers.180 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.gate_up_proj_weight_quantizers.181 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.gate_up_proj_weight_quantizers.182 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.gate_up_proj_weight_quantizers.183 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.gate_up_proj_weight_quantizers.184 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.gate_up_proj_weight_quantizers.185 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.gate_up_proj_weight_quantizers.186 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.gate_up_proj_weight_quantizers.187 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.72e-01 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.gate_up_proj_weight_quantizers.188 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.gate_up_proj_weight_quantizers.189 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.gate_up_proj_weight_quantizers.190 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.18e-01 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.gate_up_proj_weight_quantizers.191 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.19e-01 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.gate_up_proj_weight_quantizers.192 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.24e-01 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.gate_up_proj_weight_quantizers.193 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.24e-01 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.gate_up_proj_weight_quantizers.194 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.36e-01 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.gate_up_proj_weight_quantizers.195 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.gate_up_proj_weight_quantizers.196 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.gate_up_proj_weight_quantizers.197 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.23e-01 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.gate_up_proj_weight_quantizers.198 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.26e-01 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.gate_up_proj_weight_quantizers.199 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.53e-01 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.gate_up_proj_weight_quantizers.200 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.gate_up_proj_weight_quantizers.201 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.gate_up_proj_weight_quantizers.202 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.56e-01 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.gate_up_proj_weight_quantizers.203 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.gate_up_proj_weight_quantizers.204 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.gate_up_proj_weight_quantizers.205 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.50e-01 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.gate_up_proj_weight_quantizers.206 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.22e-01 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.gate_up_proj_weight_quantizers.207 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.gate_up_proj_weight_quantizers.208 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.gate_up_proj_weight_quantizers.209 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.17e-01 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.gate_up_proj_weight_quantizers.210 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.gate_up_proj_weight_quantizers.211 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.97e-01 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.gate_up_proj_weight_quantizers.212 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.gate_up_proj_weight_quantizers.213 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.29e-01 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.gate_up_proj_weight_quantizers.214 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.04e-01 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.gate_up_proj_weight_quantizers.215 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.21e-01 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.gate_up_proj_weight_quantizers.216 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.gate_up_proj_weight_quantizers.217 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.gate_up_proj_weight_quantizers.218 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.48e-01 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.gate_up_proj_weight_quantizers.219 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.gate_up_proj_weight_quantizers.220 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.gate_up_proj_weight_quantizers.221 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.45e-01 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.gate_up_proj_weight_quantizers.222 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.79e-01 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.gate_up_proj_weight_quantizers.223 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.73e-01 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.gate_up_proj_weight_quantizers.224 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.66e-01 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.gate_up_proj_weight_quantizers.225 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.gate_up_proj_weight_quantizers.226 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.gate_up_proj_weight_quantizers.227 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.85e-01 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.gate_up_proj_weight_quantizers.228 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.gate_up_proj_weight_quantizers.229 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.gate_up_proj_weight_quantizers.230 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.gate_up_proj_weight_quantizers.231 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.87e-01 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.gate_up_proj_weight_quantizers.232 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.gate_up_proj_weight_quantizers.233 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.gate_up_proj_weight_quantizers.234 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.59e-01 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.gate_up_proj_weight_quantizers.235 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.32e-01 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.gate_up_proj_weight_quantizers.236 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.77e-01 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.gate_up_proj_weight_quantizers.237 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.71e-01 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.gate_up_proj_weight_quantizers.238 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.gate_up_proj_weight_quantizers.239 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.69e-01 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.gate_up_proj_weight_quantizers.240 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=5.62e-01 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.gate_up_proj_weight_quantizers.241 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.19e-01 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.gate_up_proj_weight_quantizers.242 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.gate_up_proj_weight_quantizers.243 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.gate_up_proj_weight_quantizers.244 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.14e-01 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.gate_up_proj_weight_quantizers.245 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.gate_up_proj_weight_quantizers.246 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.gate_up_proj_weight_quantizers.247 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.17e-01 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.gate_up_proj_weight_quantizers.248 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.gate_up_proj_weight_quantizers.249 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.gate_up_proj_weight_quantizers.250 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.gate_up_proj_weight_quantizers.251 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.gate_up_proj_weight_quantizers.252 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.gate_up_proj_weight_quantizers.253 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.gate_up_proj_weight_quantizers.254 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.gate_up_proj_weight_quantizers.255 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.down_proj_input_quantizer TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.39e+01 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.down_proj_weight_quantizers.0 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.down_proj_weight_quantizers.1 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.30e-02 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.down_proj_weight_quantizers.2 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.down_proj_weight_quantizers.3 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.down_proj_weight_quantizers.4 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.64e-02 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.down_proj_weight_quantizers.5 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.91e-01 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.down_proj_weight_quantizers.6 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.59e-02 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.down_proj_weight_quantizers.7 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.down_proj_weight_quantizers.8 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.77e-01 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.down_proj_weight_quantizers.9 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.40e-02 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.down_proj_weight_quantizers.10 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.94e-01 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.down_proj_weight_quantizers.11 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.54e-02 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.down_proj_weight_quantizers.12 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.down_proj_weight_quantizers.13 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.88e-01 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.down_proj_weight_quantizers.14 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.84e-01 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.down_proj_weight_quantizers.15 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.down_proj_weight_quantizers.16 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.down_proj_weight_quantizers.17 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.down_proj_weight_quantizers.18 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.down_proj_weight_quantizers.19 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.54e-02 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.down_proj_weight_quantizers.20 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.down_proj_weight_quantizers.21 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.64e-02 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.down_proj_weight_quantizers.22 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.down_proj_weight_quantizers.23 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.24e-01 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.down_proj_weight_quantizers.24 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.down_proj_weight_quantizers.25 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.down_proj_weight_quantizers.26 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.36e-01 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.down_proj_weight_quantizers.27 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.down_proj_weight_quantizers.28 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.64e-02 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.down_proj_weight_quantizers.29 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.27e-01 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.down_proj_weight_quantizers.30 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.19e-01 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.down_proj_weight_quantizers.31 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.down_proj_weight_quantizers.32 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.down_proj_weight_quantizers.33 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.36e-01 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.down_proj_weight_quantizers.34 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.down_proj_weight_quantizers.35 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.down_proj_weight_quantizers.36 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.down_proj_weight_quantizers.37 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.down_proj_weight_quantizers.38 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.down_proj_weight_quantizers.39 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.down_proj_weight_quantizers.40 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.down_proj_weight_quantizers.41 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.39e-01 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.down_proj_weight_quantizers.42 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.down_proj_weight_quantizers.43 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.down_proj_weight_quantizers.44 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.35e-01 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.down_proj_weight_quantizers.45 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.down_proj_weight_quantizers.46 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.25e-02 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.down_proj_weight_quantizers.47 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.down_proj_weight_quantizers.48 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.50e-01 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.down_proj_weight_quantizers.49 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.59e-02 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.down_proj_weight_quantizers.50 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.23e-01 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.down_proj_weight_quantizers.51 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.down_proj_weight_quantizers.52 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.40e-02 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.down_proj_weight_quantizers.53 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.13e-01 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.down_proj_weight_quantizers.54 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.down_proj_weight_quantizers.55 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.down_proj_weight_quantizers.56 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.64e-02 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.down_proj_weight_quantizers.57 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.down_proj_weight_quantizers.58 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.18e-01 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.down_proj_weight_quantizers.59 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.down_proj_weight_quantizers.60 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.down_proj_weight_quantizers.61 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.59e-02 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.down_proj_weight_quantizers.62 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.down_proj_weight_quantizers.63 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.45e-01 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.down_proj_weight_quantizers.64 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.down_proj_weight_quantizers.65 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.54e-02 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.down_proj_weight_quantizers.66 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.68e-01 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.down_proj_weight_quantizers.67 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.down_proj_weight_quantizers.68 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.down_proj_weight_quantizers.69 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.00e-01 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.down_proj_weight_quantizers.70 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.down_proj_weight_quantizers.71 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.50e-02 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.down_proj_weight_quantizers.72 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.64e-02 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.down_proj_weight_quantizers.73 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.down_proj_weight_quantizers.74 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.down_proj_weight_quantizers.75 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.down_proj_weight_quantizers.76 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.down_proj_weight_quantizers.77 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.50e-02 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.down_proj_weight_quantizers.78 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.54e-02 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.down_proj_weight_quantizers.79 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.down_proj_weight_quantizers.80 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.down_proj_weight_quantizers.81 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.down_proj_weight_quantizers.82 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.down_proj_weight_quantizers.83 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.down_proj_weight_quantizers.84 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.38e-01 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.down_proj_weight_quantizers.85 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.54e-02 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.down_proj_weight_quantizers.86 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.51e-01 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.down_proj_weight_quantizers.87 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.down_proj_weight_quantizers.88 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.30e-02 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.down_proj_weight_quantizers.89 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.54e-02 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.down_proj_weight_quantizers.90 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.down_proj_weight_quantizers.91 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.down_proj_weight_quantizers.92 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.97e-01 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.down_proj_weight_quantizers.93 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.59e-02 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.down_proj_weight_quantizers.94 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.down_proj_weight_quantizers.95 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.03e-01 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.down_proj_weight_quantizers.96 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.down_proj_weight_quantizers.97 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.down_proj_weight_quantizers.98 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.down_proj_weight_quantizers.99 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.down_proj_weight_quantizers.100 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.down_proj_weight_quantizers.101 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.down_proj_weight_quantizers.102 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.down_proj_weight_quantizers.103 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.64e-02 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.down_proj_weight_quantizers.104 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.down_proj_weight_quantizers.105 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.down_proj_weight_quantizers.106 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.28e-01 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.down_proj_weight_quantizers.107 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.down_proj_weight_quantizers.108 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.65e-01 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.down_proj_weight_quantizers.109 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.down_proj_weight_quantizers.110 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.93e-01 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.down_proj_weight_quantizers.111 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.down_proj_weight_quantizers.112 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.64e-02 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.down_proj_weight_quantizers.113 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.49e-01 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.down_proj_weight_quantizers.114 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.down_proj_weight_quantizers.115 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.down_proj_weight_quantizers.116 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.down_proj_weight_quantizers.117 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.17e-01 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.down_proj_weight_quantizers.118 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.19e-01 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.down_proj_weight_quantizers.119 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.down_proj_weight_quantizers.120 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.62e-01 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.down_proj_weight_quantizers.121 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.down_proj_weight_quantizers.122 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.down_proj_weight_quantizers.123 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.down_proj_weight_quantizers.124 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.14e-01 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.down_proj_weight_quantizers.125 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.45e-02 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.down_proj_weight_quantizers.126 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.down_proj_weight_quantizers.127 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.down_proj_weight_quantizers.128 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.down_proj_weight_quantizers.129 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.down_proj_weight_quantizers.130 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.down_proj_weight_quantizers.131 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.34e-01 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.down_proj_weight_quantizers.132 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.down_proj_weight_quantizers.133 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.down_proj_weight_quantizers.134 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.down_proj_weight_quantizers.135 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.37e-01 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.down_proj_weight_quantizers.136 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.down_proj_weight_quantizers.137 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.30e-01 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.down_proj_weight_quantizers.138 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.02e-01 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.down_proj_weight_quantizers.139 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.96e-01 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.down_proj_weight_quantizers.140 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.down_proj_weight_quantizers.141 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.down_proj_weight_quantizers.142 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.down_proj_weight_quantizers.143 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.down_proj_weight_quantizers.144 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.66e-01 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.down_proj_weight_quantizers.145 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.down_proj_weight_quantizers.146 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.down_proj_weight_quantizers.147 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.20e-01 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.down_proj_weight_quantizers.148 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.down_proj_weight_quantizers.149 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.down_proj_weight_quantizers.150 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.down_proj_weight_quantizers.151 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.59e-02 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.down_proj_weight_quantizers.152 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.down_proj_weight_quantizers.153 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.down_proj_weight_quantizers.154 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.down_proj_weight_quantizers.155 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.down_proj_weight_quantizers.156 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.59e-02 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.down_proj_weight_quantizers.157 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.down_proj_weight_quantizers.158 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.down_proj_weight_quantizers.159 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.down_proj_weight_quantizers.160 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.14e-01 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.down_proj_weight_quantizers.161 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.down_proj_weight_quantizers.162 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.down_proj_weight_quantizers.163 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.down_proj_weight_quantizers.164 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.down_proj_weight_quantizers.165 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.50e-02 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.down_proj_weight_quantizers.166 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.54e-02 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.down_proj_weight_quantizers.167 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.45e-01 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.down_proj_weight_quantizers.168 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.down_proj_weight_quantizers.169 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.down_proj_weight_quantizers.170 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.down_proj_weight_quantizers.171 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.down_proj_weight_quantizers.172 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.down_proj_weight_quantizers.173 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.down_proj_weight_quantizers.174 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.down_proj_weight_quantizers.175 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.down_proj_weight_quantizers.176 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.down_proj_weight_quantizers.177 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.19e-01 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.down_proj_weight_quantizers.178 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.down_proj_weight_quantizers.179 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.down_proj_weight_quantizers.180 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.down_proj_weight_quantizers.181 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.down_proj_weight_quantizers.182 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.down_proj_weight_quantizers.183 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.43e-01 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.down_proj_weight_quantizers.184 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.down_proj_weight_quantizers.185 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.down_proj_weight_quantizers.186 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.down_proj_weight_quantizers.187 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.09e-01 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.down_proj_weight_quantizers.188 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.51e-01 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.down_proj_weight_quantizers.189 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.59e-02 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.down_proj_weight_quantizers.190 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.15e-01 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.down_proj_weight_quantizers.191 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.down_proj_weight_quantizers.192 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.down_proj_weight_quantizers.193 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.down_proj_weight_quantizers.194 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.down_proj_weight_quantizers.195 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.down_proj_weight_quantizers.196 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.down_proj_weight_quantizers.197 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.99e-01 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.down_proj_weight_quantizers.198 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.29e-01 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.down_proj_weight_quantizers.199 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.down_proj_weight_quantizers.200 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.85e-01 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.down_proj_weight_quantizers.201 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.down_proj_weight_quantizers.202 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.23e-01 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.down_proj_weight_quantizers.203 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.down_proj_weight_quantizers.204 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.down_proj_weight_quantizers.205 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.down_proj_weight_quantizers.206 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.37e-01 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.down_proj_weight_quantizers.207 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.down_proj_weight_quantizers.208 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.down_proj_weight_quantizers.209 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.down_proj_weight_quantizers.210 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.down_proj_weight_quantizers.211 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.down_proj_weight_quantizers.212 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.down_proj_weight_quantizers.213 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.down_proj_weight_quantizers.214 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.68e-01 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.down_proj_weight_quantizers.215 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.down_proj_weight_quantizers.216 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.down_proj_weight_quantizers.217 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.down_proj_weight_quantizers.218 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.53e-01 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.down_proj_weight_quantizers.219 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.down_proj_weight_quantizers.220 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.down_proj_weight_quantizers.221 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.down_proj_weight_quantizers.222 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.14e-01 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.down_proj_weight_quantizers.223 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.34e-01 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.down_proj_weight_quantizers.224 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.down_proj_weight_quantizers.225 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.down_proj_weight_quantizers.226 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.down_proj_weight_quantizers.227 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.down_proj_weight_quantizers.228 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.down_proj_weight_quantizers.229 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.down_proj_weight_quantizers.230 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.45e-01 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.down_proj_weight_quantizers.231 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.down_proj_weight_quantizers.232 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.down_proj_weight_quantizers.233 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.down_proj_weight_quantizers.234 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.08e-01 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.down_proj_weight_quantizers.235 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.79e-01 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.down_proj_weight_quantizers.236 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.54e-02 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.down_proj_weight_quantizers.237 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.down_proj_weight_quantizers.238 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.69e-01 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.down_proj_weight_quantizers.239 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.35e-01 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.down_proj_weight_quantizers.240 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=4.02e-01 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.down_proj_weight_quantizers.241 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.down_proj_weight_quantizers.242 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.down_proj_weight_quantizers.243 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.down_proj_weight_quantizers.244 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.down_proj_weight_quantizers.245 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.down_proj_weight_quantizers.246 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.down_proj_weight_quantizers.247 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.23e-01 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.down_proj_weight_quantizers.248 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.down_proj_weight_quantizers.249 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.down_proj_weight_quantizers.250 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.down_proj_weight_quantizers.251 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.54e-02 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.down_proj_weight_quantizers.252 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.down_proj_weight_quantizers.253 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.54e-02 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.down_proj_weight_quantizers.254 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.down_proj_weight_quantizers.255 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.32.mlp.shared_experts.gate_proj.input_quantizer TensorQuantizer(disabled) +model.layers.32.mlp.shared_experts.gate_proj.output_quantizer TensorQuantizer(disabled) +model.layers.32.mlp.shared_experts.gate_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.32.mlp.shared_experts.up_proj.input_quantizer TensorQuantizer(disabled) +model.layers.32.mlp.shared_experts.up_proj.output_quantizer TensorQuantizer(disabled) +model.layers.32.mlp.shared_experts.up_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.32.mlp.shared_experts.down_proj.input_quantizer TensorQuantizer(disabled) +model.layers.32.mlp.shared_experts.down_proj.output_quantizer TensorQuantizer(disabled) +model.layers.32.mlp.shared_experts.down_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.33.self_attn.q_a_proj.input_quantizer TensorQuantizer(disabled) +model.layers.33.self_attn.q_a_proj.output_quantizer TensorQuantizer(disabled) +model.layers.33.self_attn.q_a_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.33.self_attn.q_b_proj.input_quantizer TensorQuantizer(disabled) +model.layers.33.self_attn.q_b_proj.output_quantizer TensorQuantizer(disabled) +model.layers.33.self_attn.q_b_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.33.self_attn.kv_a_proj_with_mqa.input_quantizer TensorQuantizer(disabled) +model.layers.33.self_attn.kv_a_proj_with_mqa.output_quantizer TensorQuantizer(disabled) +model.layers.33.self_attn.kv_a_proj_with_mqa.weight_quantizer TensorQuantizer(disabled) +model.layers.33.self_attn.kv_b_proj.input_quantizer TensorQuantizer(disabled) +model.layers.33.self_attn.kv_b_proj.output_quantizer TensorQuantizer(disabled) +model.layers.33.self_attn.kv_b_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.33.self_attn.o_proj.input_quantizer TensorQuantizer(disabled) +model.layers.33.self_attn.o_proj.output_quantizer TensorQuantizer(disabled) +model.layers.33.self_attn.o_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.33.self_attn.q_bmm_quantizer TensorQuantizer(disabled) +model.layers.33.self_attn.k_bmm_quantizer TensorQuantizer((4, 3) bit fake per-tensor amax=4.48e+02(const) calibrator=MaxCalibrator quant) +model.layers.33.self_attn.v_bmm_quantizer TensorQuantizer((4, 3) bit fake per-tensor amax=4.48e+02(const) calibrator=MaxCalibrator quant) +model.layers.33.self_attn.p_bmm_quantizer TensorQuantizer(disabled) +model.layers.33.mlp.experts.gate_up_proj_input_quantizer TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.34e+00 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.gate_up_proj_weight_quantizers.0 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.40e-01 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.gate_up_proj_weight_quantizers.1 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.50e-01 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.gate_up_proj_weight_quantizers.2 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.gate_up_proj_weight_quantizers.3 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.gate_up_proj_weight_quantizers.4 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.gate_up_proj_weight_quantizers.5 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.gate_up_proj_weight_quantizers.6 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.48e-01 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.gate_up_proj_weight_quantizers.7 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.gate_up_proj_weight_quantizers.8 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.gate_up_proj_weight_quantizers.9 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.gate_up_proj_weight_quantizers.10 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.gate_up_proj_weight_quantizers.11 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.gate_up_proj_weight_quantizers.12 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.gate_up_proj_weight_quantizers.13 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.gate_up_proj_weight_quantizers.14 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.gate_up_proj_weight_quantizers.15 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.20e-01 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.gate_up_proj_weight_quantizers.16 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.gate_up_proj_weight_quantizers.17 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.gate_up_proj_weight_quantizers.18 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.gate_up_proj_weight_quantizers.19 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.gate_up_proj_weight_quantizers.20 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.gate_up_proj_weight_quantizers.21 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.gate_up_proj_weight_quantizers.22 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.94e-01 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.gate_up_proj_weight_quantizers.23 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.gate_up_proj_weight_quantizers.24 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.01e-01 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.gate_up_proj_weight_quantizers.25 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.46e-01 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.gate_up_proj_weight_quantizers.26 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.gate_up_proj_weight_quantizers.27 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.gate_up_proj_weight_quantizers.28 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.gate_up_proj_weight_quantizers.29 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.14e-01 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.gate_up_proj_weight_quantizers.30 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.gate_up_proj_weight_quantizers.31 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.62e-01 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.gate_up_proj_weight_quantizers.32 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.gate_up_proj_weight_quantizers.33 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.06e-01 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.gate_up_proj_weight_quantizers.34 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.16e-01 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.gate_up_proj_weight_quantizers.35 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.gate_up_proj_weight_quantizers.36 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.gate_up_proj_weight_quantizers.37 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.33e-01 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.gate_up_proj_weight_quantizers.38 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.gate_up_proj_weight_quantizers.39 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.gate_up_proj_weight_quantizers.40 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.33e-01 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.gate_up_proj_weight_quantizers.41 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.85e-01 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.gate_up_proj_weight_quantizers.42 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.gate_up_proj_weight_quantizers.43 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.gate_up_proj_weight_quantizers.44 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.gate_up_proj_weight_quantizers.45 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.gate_up_proj_weight_quantizers.46 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.gate_up_proj_weight_quantizers.47 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.gate_up_proj_weight_quantizers.48 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.13e-01 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.gate_up_proj_weight_quantizers.49 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.gate_up_proj_weight_quantizers.50 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.65e-01 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.gate_up_proj_weight_quantizers.51 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=4.47e-01 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.gate_up_proj_weight_quantizers.52 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.gate_up_proj_weight_quantizers.53 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.36e-01 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.gate_up_proj_weight_quantizers.54 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.gate_up_proj_weight_quantizers.55 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.gate_up_proj_weight_quantizers.56 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.gate_up_proj_weight_quantizers.57 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.gate_up_proj_weight_quantizers.58 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.47e-01 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.gate_up_proj_weight_quantizers.59 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.gate_up_proj_weight_quantizers.60 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.gate_up_proj_weight_quantizers.61 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.36e-01 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.gate_up_proj_weight_quantizers.62 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.gate_up_proj_weight_quantizers.63 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.66e-01 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.gate_up_proj_weight_quantizers.64 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.gate_up_proj_weight_quantizers.65 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.gate_up_proj_weight_quantizers.66 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.gate_up_proj_weight_quantizers.67 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.38e-01 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.gate_up_proj_weight_quantizers.68 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.gate_up_proj_weight_quantizers.69 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.gate_up_proj_weight_quantizers.70 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.gate_up_proj_weight_quantizers.71 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.gate_up_proj_weight_quantizers.72 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.gate_up_proj_weight_quantizers.73 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.64e-01 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.gate_up_proj_weight_quantizers.74 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.18e-01 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.gate_up_proj_weight_quantizers.75 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.gate_up_proj_weight_quantizers.76 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.92e-01 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.gate_up_proj_weight_quantizers.77 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.gate_up_proj_weight_quantizers.78 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.gate_up_proj_weight_quantizers.79 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.gate_up_proj_weight_quantizers.80 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.gate_up_proj_weight_quantizers.81 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.gate_up_proj_weight_quantizers.82 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.60e-01 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.gate_up_proj_weight_quantizers.83 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.47e-01 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.gate_up_proj_weight_quantizers.84 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.21e-01 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.gate_up_proj_weight_quantizers.85 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.gate_up_proj_weight_quantizers.86 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.gate_up_proj_weight_quantizers.87 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.17e-01 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.gate_up_proj_weight_quantizers.88 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.72e-01 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.gate_up_proj_weight_quantizers.89 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.gate_up_proj_weight_quantizers.90 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.15e-01 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.gate_up_proj_weight_quantizers.91 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.gate_up_proj_weight_quantizers.92 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.13e-01 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.gate_up_proj_weight_quantizers.93 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.gate_up_proj_weight_quantizers.94 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.24e-01 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.gate_up_proj_weight_quantizers.95 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.32e-01 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.gate_up_proj_weight_quantizers.96 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.gate_up_proj_weight_quantizers.97 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.32e-01 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.gate_up_proj_weight_quantizers.98 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.gate_up_proj_weight_quantizers.99 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.gate_up_proj_weight_quantizers.100 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.gate_up_proj_weight_quantizers.101 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.32e-01 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.gate_up_proj_weight_quantizers.102 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.gate_up_proj_weight_quantizers.103 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.30e-01 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.gate_up_proj_weight_quantizers.104 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.15e-01 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.gate_up_proj_weight_quantizers.105 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.gate_up_proj_weight_quantizers.106 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.gate_up_proj_weight_quantizers.107 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.gate_up_proj_weight_quantizers.108 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.gate_up_proj_weight_quantizers.109 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.gate_up_proj_weight_quantizers.110 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.gate_up_proj_weight_quantizers.111 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.58e-01 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.gate_up_proj_weight_quantizers.112 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.gate_up_proj_weight_quantizers.113 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.gate_up_proj_weight_quantizers.114 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.gate_up_proj_weight_quantizers.115 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.gate_up_proj_weight_quantizers.116 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.gate_up_proj_weight_quantizers.117 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.gate_up_proj_weight_quantizers.118 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.19e-01 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.gate_up_proj_weight_quantizers.119 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.03e-01 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.gate_up_proj_weight_quantizers.120 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.gate_up_proj_weight_quantizers.121 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.91e-01 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.gate_up_proj_weight_quantizers.122 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.gate_up_proj_weight_quantizers.123 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.gate_up_proj_weight_quantizers.124 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.gate_up_proj_weight_quantizers.125 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.gate_up_proj_weight_quantizers.126 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.19e-01 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.gate_up_proj_weight_quantizers.127 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.gate_up_proj_weight_quantizers.128 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.gate_up_proj_weight_quantizers.129 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.gate_up_proj_weight_quantizers.130 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.gate_up_proj_weight_quantizers.131 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.gate_up_proj_weight_quantizers.132 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.gate_up_proj_weight_quantizers.133 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.gate_up_proj_weight_quantizers.134 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.03e-01 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.gate_up_proj_weight_quantizers.135 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.17e-01 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.gate_up_proj_weight_quantizers.136 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.60e-01 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.gate_up_proj_weight_quantizers.137 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.gate_up_proj_weight_quantizers.138 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.gate_up_proj_weight_quantizers.139 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.gate_up_proj_weight_quantizers.140 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.gate_up_proj_weight_quantizers.141 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.gate_up_proj_weight_quantizers.142 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.33e-01 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.gate_up_proj_weight_quantizers.143 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.gate_up_proj_weight_quantizers.144 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.gate_up_proj_weight_quantizers.145 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.35e-01 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.gate_up_proj_weight_quantizers.146 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.gate_up_proj_weight_quantizers.147 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.95e-01 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.gate_up_proj_weight_quantizers.148 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.gate_up_proj_weight_quantizers.149 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.gate_up_proj_weight_quantizers.150 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.24e-01 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.gate_up_proj_weight_quantizers.151 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.gate_up_proj_weight_quantizers.152 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.gate_up_proj_weight_quantizers.153 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.gate_up_proj_weight_quantizers.154 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.59e-01 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.gate_up_proj_weight_quantizers.155 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.gate_up_proj_weight_quantizers.156 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.99e-01 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.gate_up_proj_weight_quantizers.157 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.14e-01 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.gate_up_proj_weight_quantizers.158 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.31e-01 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.gate_up_proj_weight_quantizers.159 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.gate_up_proj_weight_quantizers.160 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.gate_up_proj_weight_quantizers.161 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.gate_up_proj_weight_quantizers.162 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.gate_up_proj_weight_quantizers.163 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.32e-01 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.gate_up_proj_weight_quantizers.164 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.18e-01 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.gate_up_proj_weight_quantizers.165 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.gate_up_proj_weight_quantizers.166 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.gate_up_proj_weight_quantizers.167 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.60e-01 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.gate_up_proj_weight_quantizers.168 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.gate_up_proj_weight_quantizers.169 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.gate_up_proj_weight_quantizers.170 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.gate_up_proj_weight_quantizers.171 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.gate_up_proj_weight_quantizers.172 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.88e-01 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.gate_up_proj_weight_quantizers.173 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.gate_up_proj_weight_quantizers.174 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.gate_up_proj_weight_quantizers.175 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.19e-01 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.gate_up_proj_weight_quantizers.176 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.16e-01 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.gate_up_proj_weight_quantizers.177 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.gate_up_proj_weight_quantizers.178 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.gate_up_proj_weight_quantizers.179 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.gate_up_proj_weight_quantizers.180 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.gate_up_proj_weight_quantizers.181 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.gate_up_proj_weight_quantizers.182 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.gate_up_proj_weight_quantizers.183 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.77e-01 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.gate_up_proj_weight_quantizers.184 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.gate_up_proj_weight_quantizers.185 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.gate_up_proj_weight_quantizers.186 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.gate_up_proj_weight_quantizers.187 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.gate_up_proj_weight_quantizers.188 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.84e-01 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.gate_up_proj_weight_quantizers.189 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.58e-01 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.gate_up_proj_weight_quantizers.190 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.gate_up_proj_weight_quantizers.191 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.13e-01 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.gate_up_proj_weight_quantizers.192 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.gate_up_proj_weight_quantizers.193 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.gate_up_proj_weight_quantizers.194 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.gate_up_proj_weight_quantizers.195 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.75e-01 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.gate_up_proj_weight_quantizers.196 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.34e-01 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.gate_up_proj_weight_quantizers.197 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.gate_up_proj_weight_quantizers.198 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.60e-01 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.gate_up_proj_weight_quantizers.199 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.20e-01 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.gate_up_proj_weight_quantizers.200 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.gate_up_proj_weight_quantizers.201 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.gate_up_proj_weight_quantizers.202 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.gate_up_proj_weight_quantizers.203 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.gate_up_proj_weight_quantizers.204 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.gate_up_proj_weight_quantizers.205 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.31e-01 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.gate_up_proj_weight_quantizers.206 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.26e-01 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.gate_up_proj_weight_quantizers.207 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.47e-01 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.gate_up_proj_weight_quantizers.208 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.gate_up_proj_weight_quantizers.209 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.gate_up_proj_weight_quantizers.210 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.69e-01 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.gate_up_proj_weight_quantizers.211 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.53e-01 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.gate_up_proj_weight_quantizers.212 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.34e-01 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.gate_up_proj_weight_quantizers.213 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.gate_up_proj_weight_quantizers.214 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.18e-01 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.gate_up_proj_weight_quantizers.215 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.gate_up_proj_weight_quantizers.216 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.gate_up_proj_weight_quantizers.217 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.gate_up_proj_weight_quantizers.218 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.40e-01 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.gate_up_proj_weight_quantizers.219 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.gate_up_proj_weight_quantizers.220 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.22e-01 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.gate_up_proj_weight_quantizers.221 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.08e-01 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.gate_up_proj_weight_quantizers.222 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.38e-01 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.gate_up_proj_weight_quantizers.223 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.31e-01 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.gate_up_proj_weight_quantizers.224 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.gate_up_proj_weight_quantizers.225 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.31e-01 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.gate_up_proj_weight_quantizers.226 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.39e-01 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.gate_up_proj_weight_quantizers.227 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.gate_up_proj_weight_quantizers.228 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.gate_up_proj_weight_quantizers.229 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.gate_up_proj_weight_quantizers.230 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.22e-01 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.gate_up_proj_weight_quantizers.231 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.15e-01 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.gate_up_proj_weight_quantizers.232 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.gate_up_proj_weight_quantizers.233 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.96e-01 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.gate_up_proj_weight_quantizers.234 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.21e-01 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.gate_up_proj_weight_quantizers.235 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.gate_up_proj_weight_quantizers.236 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.gate_up_proj_weight_quantizers.237 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.gate_up_proj_weight_quantizers.238 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.gate_up_proj_weight_quantizers.239 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.gate_up_proj_weight_quantizers.240 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.gate_up_proj_weight_quantizers.241 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.50e-01 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.gate_up_proj_weight_quantizers.242 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.gate_up_proj_weight_quantizers.243 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.11e-01 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.gate_up_proj_weight_quantizers.244 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.32e-01 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.gate_up_proj_weight_quantizers.245 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.34e-01 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.gate_up_proj_weight_quantizers.246 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.gate_up_proj_weight_quantizers.247 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.38e-01 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.gate_up_proj_weight_quantizers.248 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.gate_up_proj_weight_quantizers.249 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.91e-01 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.gate_up_proj_weight_quantizers.250 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.gate_up_proj_weight_quantizers.251 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.gate_up_proj_weight_quantizers.252 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.gate_up_proj_weight_quantizers.253 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.55e-01 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.gate_up_proj_weight_quantizers.254 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.gate_up_proj_weight_quantizers.255 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.12e-01 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.down_proj_input_quantizer TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.21e+01 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.down_proj_weight_quantizers.0 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.21e-01 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.down_proj_weight_quantizers.1 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.64e-02 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.down_proj_weight_quantizers.2 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.down_proj_weight_quantizers.3 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.down_proj_weight_quantizers.4 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.64e-02 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.down_proj_weight_quantizers.5 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.50e-02 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.down_proj_weight_quantizers.6 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.down_proj_weight_quantizers.7 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.45e-02 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.down_proj_weight_quantizers.8 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.down_proj_weight_quantizers.9 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.down_proj_weight_quantizers.10 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.54e-02 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.down_proj_weight_quantizers.11 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.down_proj_weight_quantizers.12 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.down_proj_weight_quantizers.13 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.down_proj_weight_quantizers.14 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.down_proj_weight_quantizers.15 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.down_proj_weight_quantizers.16 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.down_proj_weight_quantizers.17 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.down_proj_weight_quantizers.18 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.down_proj_weight_quantizers.19 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.down_proj_weight_quantizers.20 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.down_proj_weight_quantizers.21 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.59e-02 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.down_proj_weight_quantizers.22 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.down_proj_weight_quantizers.23 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.down_proj_weight_quantizers.24 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.down_proj_weight_quantizers.25 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.59e-02 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.down_proj_weight_quantizers.26 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.59e-02 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.down_proj_weight_quantizers.27 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.down_proj_weight_quantizers.28 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.54e-02 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.down_proj_weight_quantizers.29 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.down_proj_weight_quantizers.30 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.down_proj_weight_quantizers.31 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.down_proj_weight_quantizers.32 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.down_proj_weight_quantizers.33 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.down_proj_weight_quantizers.34 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.down_proj_weight_quantizers.35 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.down_proj_weight_quantizers.36 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.64e-02 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.down_proj_weight_quantizers.37 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.down_proj_weight_quantizers.38 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.down_proj_weight_quantizers.39 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.down_proj_weight_quantizers.40 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.down_proj_weight_quantizers.41 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.25e-01 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.down_proj_weight_quantizers.42 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.down_proj_weight_quantizers.43 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.down_proj_weight_quantizers.44 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.down_proj_weight_quantizers.45 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.down_proj_weight_quantizers.46 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.down_proj_weight_quantizers.47 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.down_proj_weight_quantizers.48 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.down_proj_weight_quantizers.49 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.down_proj_weight_quantizers.50 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.32e-01 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.down_proj_weight_quantizers.51 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.65e-01 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.down_proj_weight_quantizers.52 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.down_proj_weight_quantizers.53 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.down_proj_weight_quantizers.54 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.down_proj_weight_quantizers.55 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.down_proj_weight_quantizers.56 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.down_proj_weight_quantizers.57 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.down_proj_weight_quantizers.58 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.46e-01 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.down_proj_weight_quantizers.59 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.down_proj_weight_quantizers.60 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.45e-02 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.down_proj_weight_quantizers.61 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.down_proj_weight_quantizers.62 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.down_proj_weight_quantizers.63 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.54e-02 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.down_proj_weight_quantizers.64 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.down_proj_weight_quantizers.65 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.down_proj_weight_quantizers.66 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.down_proj_weight_quantizers.67 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.59e-02 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.down_proj_weight_quantizers.68 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.down_proj_weight_quantizers.69 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.64e-02 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.down_proj_weight_quantizers.70 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.down_proj_weight_quantizers.71 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.54e-02 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.down_proj_weight_quantizers.72 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.down_proj_weight_quantizers.73 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.50e-02 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.down_proj_weight_quantizers.74 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.down_proj_weight_quantizers.75 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.down_proj_weight_quantizers.76 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.55e-01 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.down_proj_weight_quantizers.77 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.down_proj_weight_quantizers.78 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.down_proj_weight_quantizers.79 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.54e-02 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.down_proj_weight_quantizers.80 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.down_proj_weight_quantizers.81 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.30e-02 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.down_proj_weight_quantizers.82 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.63e-01 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.down_proj_weight_quantizers.83 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.59e-02 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.down_proj_weight_quantizers.84 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.down_proj_weight_quantizers.85 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.25e-02 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.down_proj_weight_quantizers.86 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.down_proj_weight_quantizers.87 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.40e-01 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.down_proj_weight_quantizers.88 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.down_proj_weight_quantizers.89 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.down_proj_weight_quantizers.90 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.down_proj_weight_quantizers.91 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.54e-02 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.down_proj_weight_quantizers.92 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.down_proj_weight_quantizers.93 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.down_proj_weight_quantizers.94 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.down_proj_weight_quantizers.95 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.38e-01 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.down_proj_weight_quantizers.96 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.down_proj_weight_quantizers.97 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.44e-01 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.down_proj_weight_quantizers.98 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.down_proj_weight_quantizers.99 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.down_proj_weight_quantizers.100 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.down_proj_weight_quantizers.101 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.30e-02 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.down_proj_weight_quantizers.102 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.down_proj_weight_quantizers.103 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.down_proj_weight_quantizers.104 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.22e-01 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.down_proj_weight_quantizers.105 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.down_proj_weight_quantizers.106 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.40e-02 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.down_proj_weight_quantizers.107 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.down_proj_weight_quantizers.108 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.down_proj_weight_quantizers.109 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.down_proj_weight_quantizers.110 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.14e-01 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.down_proj_weight_quantizers.111 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.56e-01 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.down_proj_weight_quantizers.112 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.down_proj_weight_quantizers.113 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.down_proj_weight_quantizers.114 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.64e-02 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.down_proj_weight_quantizers.115 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.down_proj_weight_quantizers.116 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.down_proj_weight_quantizers.117 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.11e-02 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.down_proj_weight_quantizers.118 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.down_proj_weight_quantizers.119 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.13e-01 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.down_proj_weight_quantizers.120 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.50e-02 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.down_proj_weight_quantizers.121 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.26e-01 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.down_proj_weight_quantizers.122 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.down_proj_weight_quantizers.123 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.down_proj_weight_quantizers.124 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.down_proj_weight_quantizers.125 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.down_proj_weight_quantizers.126 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.down_proj_weight_quantizers.127 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.down_proj_weight_quantizers.128 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.down_proj_weight_quantizers.129 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.down_proj_weight_quantizers.130 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.down_proj_weight_quantizers.131 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.40e-02 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.down_proj_weight_quantizers.132 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.down_proj_weight_quantizers.133 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.45e-02 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.down_proj_weight_quantizers.134 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.88e-01 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.down_proj_weight_quantizers.135 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.down_proj_weight_quantizers.136 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.down_proj_weight_quantizers.137 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.down_proj_weight_quantizers.138 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.down_proj_weight_quantizers.139 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.down_proj_weight_quantizers.140 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.down_proj_weight_quantizers.141 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.50e-02 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.down_proj_weight_quantizers.142 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.down_proj_weight_quantizers.143 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.43e-01 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.down_proj_weight_quantizers.144 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.35e-02 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.down_proj_weight_quantizers.145 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.down_proj_weight_quantizers.146 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.down_proj_weight_quantizers.147 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.20e-01 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.down_proj_weight_quantizers.148 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.59e-02 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.down_proj_weight_quantizers.149 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.down_proj_weight_quantizers.150 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.50e-02 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.down_proj_weight_quantizers.151 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.down_proj_weight_quantizers.152 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.down_proj_weight_quantizers.153 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.down_proj_weight_quantizers.154 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.48e-01 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.down_proj_weight_quantizers.155 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.59e-02 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.down_proj_weight_quantizers.156 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.70e-01 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.down_proj_weight_quantizers.157 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.down_proj_weight_quantizers.158 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.30e-01 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.down_proj_weight_quantizers.159 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.down_proj_weight_quantizers.160 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.down_proj_weight_quantizers.161 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.down_proj_weight_quantizers.162 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.down_proj_weight_quantizers.163 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.down_proj_weight_quantizers.164 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.59e-02 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.down_proj_weight_quantizers.165 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.down_proj_weight_quantizers.166 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.down_proj_weight_quantizers.167 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.62e-01 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.down_proj_weight_quantizers.168 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.down_proj_weight_quantizers.169 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.down_proj_weight_quantizers.170 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.down_proj_weight_quantizers.171 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.down_proj_weight_quantizers.172 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.down_proj_weight_quantizers.173 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.down_proj_weight_quantizers.174 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.down_proj_weight_quantizers.175 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.down_proj_weight_quantizers.176 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.down_proj_weight_quantizers.177 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.45e-02 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.down_proj_weight_quantizers.178 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.down_proj_weight_quantizers.179 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.down_proj_weight_quantizers.180 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.down_proj_weight_quantizers.181 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.down_proj_weight_quantizers.182 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.down_proj_weight_quantizers.183 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.down_proj_weight_quantizers.184 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.down_proj_weight_quantizers.185 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.down_proj_weight_quantizers.186 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.down_proj_weight_quantizers.187 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.down_proj_weight_quantizers.188 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.down_proj_weight_quantizers.189 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.25e-01 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.down_proj_weight_quantizers.190 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.down_proj_weight_quantizers.191 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.down_proj_weight_quantizers.192 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.down_proj_weight_quantizers.193 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.down_proj_weight_quantizers.194 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.40e-02 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.down_proj_weight_quantizers.195 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.38e-01 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.down_proj_weight_quantizers.196 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.down_proj_weight_quantizers.197 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.down_proj_weight_quantizers.198 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.60e-01 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.down_proj_weight_quantizers.199 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.down_proj_weight_quantizers.200 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.down_proj_weight_quantizers.201 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.64e-02 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.down_proj_weight_quantizers.202 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.down_proj_weight_quantizers.203 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.down_proj_weight_quantizers.204 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.down_proj_weight_quantizers.205 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.down_proj_weight_quantizers.206 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.down_proj_weight_quantizers.207 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.down_proj_weight_quantizers.208 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.down_proj_weight_quantizers.209 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.down_proj_weight_quantizers.210 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.27e-01 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.down_proj_weight_quantizers.211 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.down_proj_weight_quantizers.212 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.21e-01 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.down_proj_weight_quantizers.213 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.50e-02 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.down_proj_weight_quantizers.214 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.down_proj_weight_quantizers.215 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.down_proj_weight_quantizers.216 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.down_proj_weight_quantizers.217 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.down_proj_weight_quantizers.218 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.down_proj_weight_quantizers.219 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.50e-02 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.down_proj_weight_quantizers.220 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.13e-01 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.down_proj_weight_quantizers.221 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.down_proj_weight_quantizers.222 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.down_proj_weight_quantizers.223 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.down_proj_weight_quantizers.224 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.down_proj_weight_quantizers.225 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.down_proj_weight_quantizers.226 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.down_proj_weight_quantizers.227 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.down_proj_weight_quantizers.228 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.down_proj_weight_quantizers.229 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.35e-02 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.down_proj_weight_quantizers.230 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.down_proj_weight_quantizers.231 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.down_proj_weight_quantizers.232 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.64e-02 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.down_proj_weight_quantizers.233 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.down_proj_weight_quantizers.234 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.91e-01 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.down_proj_weight_quantizers.235 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.down_proj_weight_quantizers.236 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.down_proj_weight_quantizers.237 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.down_proj_weight_quantizers.238 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.down_proj_weight_quantizers.239 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.down_proj_weight_quantizers.240 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.down_proj_weight_quantizers.241 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.down_proj_weight_quantizers.242 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.54e-02 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.down_proj_weight_quantizers.243 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.down_proj_weight_quantizers.244 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.64e-02 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.down_proj_weight_quantizers.245 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.68e-01 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.down_proj_weight_quantizers.246 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.down_proj_weight_quantizers.247 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.down_proj_weight_quantizers.248 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.down_proj_weight_quantizers.249 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.18e-01 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.down_proj_weight_quantizers.250 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.down_proj_weight_quantizers.251 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.down_proj_weight_quantizers.252 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.down_proj_weight_quantizers.253 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.down_proj_weight_quantizers.254 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.down_proj_weight_quantizers.255 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.33.mlp.shared_experts.gate_proj.input_quantizer TensorQuantizer(disabled) +model.layers.33.mlp.shared_experts.gate_proj.output_quantizer TensorQuantizer(disabled) +model.layers.33.mlp.shared_experts.gate_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.33.mlp.shared_experts.up_proj.input_quantizer TensorQuantizer(disabled) +model.layers.33.mlp.shared_experts.up_proj.output_quantizer TensorQuantizer(disabled) +model.layers.33.mlp.shared_experts.up_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.33.mlp.shared_experts.down_proj.input_quantizer TensorQuantizer(disabled) +model.layers.33.mlp.shared_experts.down_proj.output_quantizer TensorQuantizer(disabled) +model.layers.33.mlp.shared_experts.down_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.34.self_attn.q_a_proj.input_quantizer TensorQuantizer(disabled) +model.layers.34.self_attn.q_a_proj.output_quantizer TensorQuantizer(disabled) +model.layers.34.self_attn.q_a_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.34.self_attn.q_b_proj.input_quantizer TensorQuantizer(disabled) +model.layers.34.self_attn.q_b_proj.output_quantizer TensorQuantizer(disabled) +model.layers.34.self_attn.q_b_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.34.self_attn.kv_a_proj_with_mqa.input_quantizer TensorQuantizer(disabled) +model.layers.34.self_attn.kv_a_proj_with_mqa.output_quantizer TensorQuantizer(disabled) +model.layers.34.self_attn.kv_a_proj_with_mqa.weight_quantizer TensorQuantizer(disabled) +model.layers.34.self_attn.kv_b_proj.input_quantizer TensorQuantizer(disabled) +model.layers.34.self_attn.kv_b_proj.output_quantizer TensorQuantizer(disabled) +model.layers.34.self_attn.kv_b_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.34.self_attn.o_proj.input_quantizer TensorQuantizer(disabled) +model.layers.34.self_attn.o_proj.output_quantizer TensorQuantizer(disabled) +model.layers.34.self_attn.o_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.34.self_attn.indexer.wq_b.input_quantizer TensorQuantizer(disabled) +model.layers.34.self_attn.indexer.wq_b.output_quantizer TensorQuantizer(disabled) +model.layers.34.self_attn.indexer.wq_b.weight_quantizer TensorQuantizer(disabled) +model.layers.34.self_attn.indexer.wk.input_quantizer TensorQuantizer(disabled) +model.layers.34.self_attn.indexer.wk.output_quantizer TensorQuantizer(disabled) +model.layers.34.self_attn.indexer.wk.weight_quantizer TensorQuantizer(disabled) +model.layers.34.self_attn.indexer.k_norm.input_quantizer TensorQuantizer(disabled) +model.layers.34.self_attn.indexer.k_norm.output_quantizer TensorQuantizer(disabled) +model.layers.34.self_attn.indexer.weights_proj.input_quantizer TensorQuantizer(disabled) +model.layers.34.self_attn.indexer.weights_proj.output_quantizer TensorQuantizer(disabled) +model.layers.34.self_attn.indexer.weights_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.34.self_attn.q_bmm_quantizer TensorQuantizer(disabled) +model.layers.34.self_attn.k_bmm_quantizer TensorQuantizer((4, 3) bit fake per-tensor amax=4.48e+02(const) calibrator=MaxCalibrator quant) +model.layers.34.self_attn.v_bmm_quantizer TensorQuantizer((4, 3) bit fake per-tensor amax=4.48e+02(const) calibrator=MaxCalibrator quant) +model.layers.34.self_attn.p_bmm_quantizer TensorQuantizer(disabled) +model.layers.34.mlp.experts.gate_up_proj_input_quantizer TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.56e+00 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.gate_up_proj_weight_quantizers.0 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.gate_up_proj_weight_quantizers.1 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.97e-01 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.gate_up_proj_weight_quantizers.2 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.34e-01 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.gate_up_proj_weight_quantizers.3 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.83e-01 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.gate_up_proj_weight_quantizers.4 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.96e-01 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.gate_up_proj_weight_quantizers.5 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.gate_up_proj_weight_quantizers.6 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.35e-01 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.gate_up_proj_weight_quantizers.7 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.gate_up_proj_weight_quantizers.8 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.gate_up_proj_weight_quantizers.9 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.gate_up_proj_weight_quantizers.10 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.gate_up_proj_weight_quantizers.11 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.gate_up_proj_weight_quantizers.12 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.gate_up_proj_weight_quantizers.13 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.gate_up_proj_weight_quantizers.14 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.14e-01 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.gate_up_proj_weight_quantizers.15 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.36e-01 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.gate_up_proj_weight_quantizers.16 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.gate_up_proj_weight_quantizers.17 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.79e-01 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.gate_up_proj_weight_quantizers.18 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.50e-01 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.gate_up_proj_weight_quantizers.19 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.gate_up_proj_weight_quantizers.20 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.gate_up_proj_weight_quantizers.21 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.26e-01 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.gate_up_proj_weight_quantizers.22 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.gate_up_proj_weight_quantizers.23 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.gate_up_proj_weight_quantizers.24 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.gate_up_proj_weight_quantizers.25 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.79e-01 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.gate_up_proj_weight_quantizers.26 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.gate_up_proj_weight_quantizers.27 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.gate_up_proj_weight_quantizers.28 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=6.48e-01 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.gate_up_proj_weight_quantizers.29 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.gate_up_proj_weight_quantizers.30 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.gate_up_proj_weight_quantizers.31 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.gate_up_proj_weight_quantizers.32 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.68e-01 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.gate_up_proj_weight_quantizers.33 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.gate_up_proj_weight_quantizers.34 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.gate_up_proj_weight_quantizers.35 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.97e-01 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.gate_up_proj_weight_quantizers.36 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.gate_up_proj_weight_quantizers.37 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.22e-01 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.gate_up_proj_weight_quantizers.38 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.42e-01 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.gate_up_proj_weight_quantizers.39 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.gate_up_proj_weight_quantizers.40 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.02e-01 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.gate_up_proj_weight_quantizers.41 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.gate_up_proj_weight_quantizers.42 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.23e-01 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.gate_up_proj_weight_quantizers.43 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.gate_up_proj_weight_quantizers.44 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.15e-01 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.gate_up_proj_weight_quantizers.45 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.gate_up_proj_weight_quantizers.46 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.gate_up_proj_weight_quantizers.47 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.gate_up_proj_weight_quantizers.48 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.gate_up_proj_weight_quantizers.49 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.30e-01 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.gate_up_proj_weight_quantizers.50 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.gate_up_proj_weight_quantizers.51 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.gate_up_proj_weight_quantizers.52 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.gate_up_proj_weight_quantizers.53 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.gate_up_proj_weight_quantizers.54 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.gate_up_proj_weight_quantizers.55 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.gate_up_proj_weight_quantizers.56 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.gate_up_proj_weight_quantizers.57 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.40e-01 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.gate_up_proj_weight_quantizers.58 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.gate_up_proj_weight_quantizers.59 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.32e-01 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.gate_up_proj_weight_quantizers.60 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.gate_up_proj_weight_quantizers.61 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.13e-01 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.gate_up_proj_weight_quantizers.62 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.gate_up_proj_weight_quantizers.63 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.45e-01 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.gate_up_proj_weight_quantizers.64 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.gate_up_proj_weight_quantizers.65 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.gate_up_proj_weight_quantizers.66 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.15e-01 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.gate_up_proj_weight_quantizers.67 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.gate_up_proj_weight_quantizers.68 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.gate_up_proj_weight_quantizers.69 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.59e-02 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.gate_up_proj_weight_quantizers.70 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.gate_up_proj_weight_quantizers.71 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.19e-01 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.gate_up_proj_weight_quantizers.72 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.gate_up_proj_weight_quantizers.73 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.49e-01 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.gate_up_proj_weight_quantizers.74 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.gate_up_proj_weight_quantizers.75 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.gate_up_proj_weight_quantizers.76 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.gate_up_proj_weight_quantizers.77 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.23e-01 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.gate_up_proj_weight_quantizers.78 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.17e-01 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.gate_up_proj_weight_quantizers.79 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.gate_up_proj_weight_quantizers.80 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.20e-01 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.gate_up_proj_weight_quantizers.81 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.gate_up_proj_weight_quantizers.82 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.gate_up_proj_weight_quantizers.83 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.91e-01 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.gate_up_proj_weight_quantizers.84 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.70e-01 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.gate_up_proj_weight_quantizers.85 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.77e-01 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.gate_up_proj_weight_quantizers.86 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.22e-01 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.gate_up_proj_weight_quantizers.87 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.gate_up_proj_weight_quantizers.88 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.02e-01 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.gate_up_proj_weight_quantizers.89 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.gate_up_proj_weight_quantizers.90 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.gate_up_proj_weight_quantizers.91 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.29e-01 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.gate_up_proj_weight_quantizers.92 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.gate_up_proj_weight_quantizers.93 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.gate_up_proj_weight_quantizers.94 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.gate_up_proj_weight_quantizers.95 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.gate_up_proj_weight_quantizers.96 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.59e-01 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.gate_up_proj_weight_quantizers.97 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.gate_up_proj_weight_quantizers.98 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.gate_up_proj_weight_quantizers.99 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.gate_up_proj_weight_quantizers.100 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.gate_up_proj_weight_quantizers.101 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.gate_up_proj_weight_quantizers.102 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.20e-01 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.gate_up_proj_weight_quantizers.103 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.gate_up_proj_weight_quantizers.104 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.gate_up_proj_weight_quantizers.105 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.58e-01 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.gate_up_proj_weight_quantizers.106 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.gate_up_proj_weight_quantizers.107 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.gate_up_proj_weight_quantizers.108 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.25e-01 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.gate_up_proj_weight_quantizers.109 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.gate_up_proj_weight_quantizers.110 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.gate_up_proj_weight_quantizers.111 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.gate_up_proj_weight_quantizers.112 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.41e-01 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.gate_up_proj_weight_quantizers.113 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.gate_up_proj_weight_quantizers.114 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.gate_up_proj_weight_quantizers.115 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.56e-01 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.gate_up_proj_weight_quantizers.116 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.58e-01 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.gate_up_proj_weight_quantizers.117 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.18e-01 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.gate_up_proj_weight_quantizers.118 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.gate_up_proj_weight_quantizers.119 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.gate_up_proj_weight_quantizers.120 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.gate_up_proj_weight_quantizers.121 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.11e-01 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.gate_up_proj_weight_quantizers.122 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.gate_up_proj_weight_quantizers.123 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.gate_up_proj_weight_quantizers.124 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.gate_up_proj_weight_quantizers.125 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.gate_up_proj_weight_quantizers.126 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.gate_up_proj_weight_quantizers.127 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.29e-01 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.gate_up_proj_weight_quantizers.128 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.17e-01 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.gate_up_proj_weight_quantizers.129 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.gate_up_proj_weight_quantizers.130 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.38e-01 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.gate_up_proj_weight_quantizers.131 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.gate_up_proj_weight_quantizers.132 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.gate_up_proj_weight_quantizers.133 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.69e-01 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.gate_up_proj_weight_quantizers.134 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.08e-01 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.gate_up_proj_weight_quantizers.135 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.52e-01 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.gate_up_proj_weight_quantizers.136 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.gate_up_proj_weight_quantizers.137 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.gate_up_proj_weight_quantizers.138 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.gate_up_proj_weight_quantizers.139 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.21e-01 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.gate_up_proj_weight_quantizers.140 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.gate_up_proj_weight_quantizers.141 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.gate_up_proj_weight_quantizers.142 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.gate_up_proj_weight_quantizers.143 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.gate_up_proj_weight_quantizers.144 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.gate_up_proj_weight_quantizers.145 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.80e-01 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.gate_up_proj_weight_quantizers.146 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=4.55e-01 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.gate_up_proj_weight_quantizers.147 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.32e-01 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.gate_up_proj_weight_quantizers.148 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.gate_up_proj_weight_quantizers.149 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.16e-01 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.gate_up_proj_weight_quantizers.150 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.gate_up_proj_weight_quantizers.151 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.43e-01 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.gate_up_proj_weight_quantizers.152 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.gate_up_proj_weight_quantizers.153 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.gate_up_proj_weight_quantizers.154 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.gate_up_proj_weight_quantizers.155 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.31e-01 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.gate_up_proj_weight_quantizers.156 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.18e-01 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.gate_up_proj_weight_quantizers.157 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.gate_up_proj_weight_quantizers.158 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.91e-01 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.gate_up_proj_weight_quantizers.159 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.gate_up_proj_weight_quantizers.160 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.gate_up_proj_weight_quantizers.161 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.30e-01 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.gate_up_proj_weight_quantizers.162 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.gate_up_proj_weight_quantizers.163 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.gate_up_proj_weight_quantizers.164 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.gate_up_proj_weight_quantizers.165 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.59e-01 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.gate_up_proj_weight_quantizers.166 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.14e-01 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.gate_up_proj_weight_quantizers.167 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.13e-01 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.gate_up_proj_weight_quantizers.168 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.gate_up_proj_weight_quantizers.169 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.gate_up_proj_weight_quantizers.170 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.68e-01 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.gate_up_proj_weight_quantizers.171 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.gate_up_proj_weight_quantizers.172 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.36e-01 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.gate_up_proj_weight_quantizers.173 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.gate_up_proj_weight_quantizers.174 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.gate_up_proj_weight_quantizers.175 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.gate_up_proj_weight_quantizers.176 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.29e-01 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.gate_up_proj_weight_quantizers.177 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.30e-01 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.gate_up_proj_weight_quantizers.178 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.gate_up_proj_weight_quantizers.179 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.gate_up_proj_weight_quantizers.180 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=4.08e-01 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.gate_up_proj_weight_quantizers.181 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.03e-01 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.gate_up_proj_weight_quantizers.182 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.23e-01 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.gate_up_proj_weight_quantizers.183 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.gate_up_proj_weight_quantizers.184 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.gate_up_proj_weight_quantizers.185 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.gate_up_proj_weight_quantizers.186 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.87e-01 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.gate_up_proj_weight_quantizers.187 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.88e-01 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.gate_up_proj_weight_quantizers.188 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.45e-01 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.gate_up_proj_weight_quantizers.189 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.16e-01 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.gate_up_proj_weight_quantizers.190 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.gate_up_proj_weight_quantizers.191 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.31e-01 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.gate_up_proj_weight_quantizers.192 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.87e-01 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.gate_up_proj_weight_quantizers.193 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.gate_up_proj_weight_quantizers.194 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.gate_up_proj_weight_quantizers.195 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.gate_up_proj_weight_quantizers.196 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.gate_up_proj_weight_quantizers.197 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.64e-01 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.gate_up_proj_weight_quantizers.198 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.gate_up_proj_weight_quantizers.199 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.gate_up_proj_weight_quantizers.200 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.36e-01 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.gate_up_proj_weight_quantizers.201 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.gate_up_proj_weight_quantizers.202 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.gate_up_proj_weight_quantizers.203 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.gate_up_proj_weight_quantizers.204 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.gate_up_proj_weight_quantizers.205 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.gate_up_proj_weight_quantizers.206 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.gate_up_proj_weight_quantizers.207 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.18e-01 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.gate_up_proj_weight_quantizers.208 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.gate_up_proj_weight_quantizers.209 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.gate_up_proj_weight_quantizers.210 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.16e-01 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.gate_up_proj_weight_quantizers.211 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.09e-01 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.gate_up_proj_weight_quantizers.212 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.gate_up_proj_weight_quantizers.213 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.gate_up_proj_weight_quantizers.214 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.62e-01 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.gate_up_proj_weight_quantizers.215 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.gate_up_proj_weight_quantizers.216 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.20e-01 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.gate_up_proj_weight_quantizers.217 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.29e-01 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.gate_up_proj_weight_quantizers.218 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.gate_up_proj_weight_quantizers.219 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.56e-01 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.gate_up_proj_weight_quantizers.220 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.84e-01 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.gate_up_proj_weight_quantizers.221 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.gate_up_proj_weight_quantizers.222 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.15e-01 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.gate_up_proj_weight_quantizers.223 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.gate_up_proj_weight_quantizers.224 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.gate_up_proj_weight_quantizers.225 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.gate_up_proj_weight_quantizers.226 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.gate_up_proj_weight_quantizers.227 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.70e-01 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.gate_up_proj_weight_quantizers.228 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.32e-01 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.gate_up_proj_weight_quantizers.229 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.26e-01 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.gate_up_proj_weight_quantizers.230 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.gate_up_proj_weight_quantizers.231 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.gate_up_proj_weight_quantizers.232 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.73e-01 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.gate_up_proj_weight_quantizers.233 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.35e-01 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.gate_up_proj_weight_quantizers.234 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.36e-01 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.gate_up_proj_weight_quantizers.235 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.gate_up_proj_weight_quantizers.236 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.gate_up_proj_weight_quantizers.237 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.gate_up_proj_weight_quantizers.238 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.gate_up_proj_weight_quantizers.239 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.74e-01 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.gate_up_proj_weight_quantizers.240 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.88e-01 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.gate_up_proj_weight_quantizers.241 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.gate_up_proj_weight_quantizers.242 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.45e-01 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.gate_up_proj_weight_quantizers.243 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.gate_up_proj_weight_quantizers.244 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.49e-01 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.gate_up_proj_weight_quantizers.245 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.gate_up_proj_weight_quantizers.246 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.gate_up_proj_weight_quantizers.247 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.gate_up_proj_weight_quantizers.248 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.gate_up_proj_weight_quantizers.249 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.gate_up_proj_weight_quantizers.250 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.gate_up_proj_weight_quantizers.251 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.03e-01 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.gate_up_proj_weight_quantizers.252 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.20e-01 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.gate_up_proj_weight_quantizers.253 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.24e-01 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.gate_up_proj_weight_quantizers.254 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.gate_up_proj_weight_quantizers.255 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.down_proj_input_quantizer TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=5.02e+01 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.down_proj_weight_quantizers.0 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.down_proj_weight_quantizers.1 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.13e-01 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.down_proj_weight_quantizers.2 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.64e-02 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.down_proj_weight_quantizers.3 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.35e-01 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.down_proj_weight_quantizers.4 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.down_proj_weight_quantizers.5 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.down_proj_weight_quantizers.6 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.down_proj_weight_quantizers.7 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.down_proj_weight_quantizers.8 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.down_proj_weight_quantizers.9 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.54e-02 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.down_proj_weight_quantizers.10 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.down_proj_weight_quantizers.11 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.down_proj_weight_quantizers.12 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.down_proj_weight_quantizers.13 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.down_proj_weight_quantizers.14 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.down_proj_weight_quantizers.15 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.down_proj_weight_quantizers.16 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.39e-01 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.down_proj_weight_quantizers.17 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.21e-01 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.down_proj_weight_quantizers.18 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.down_proj_weight_quantizers.19 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.down_proj_weight_quantizers.20 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.down_proj_weight_quantizers.21 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.49e-01 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.down_proj_weight_quantizers.22 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.down_proj_weight_quantizers.23 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.down_proj_weight_quantizers.24 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.down_proj_weight_quantizers.25 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.down_proj_weight_quantizers.26 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.down_proj_weight_quantizers.27 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.down_proj_weight_quantizers.28 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.28e-01 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.down_proj_weight_quantizers.29 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.down_proj_weight_quantizers.30 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.down_proj_weight_quantizers.31 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.down_proj_weight_quantizers.32 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.down_proj_weight_quantizers.33 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.down_proj_weight_quantizers.34 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.down_proj_weight_quantizers.35 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.22e-01 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.down_proj_weight_quantizers.36 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.down_proj_weight_quantizers.37 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.down_proj_weight_quantizers.38 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.down_proj_weight_quantizers.39 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.down_proj_weight_quantizers.40 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.down_proj_weight_quantizers.41 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.down_proj_weight_quantizers.42 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.down_proj_weight_quantizers.43 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.down_proj_weight_quantizers.44 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.54e-02 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.down_proj_weight_quantizers.45 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.35e-02 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.down_proj_weight_quantizers.46 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.down_proj_weight_quantizers.47 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.64e-02 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.down_proj_weight_quantizers.48 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.64e-02 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.down_proj_weight_quantizers.49 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.down_proj_weight_quantizers.50 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.59e-02 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.down_proj_weight_quantizers.51 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.down_proj_weight_quantizers.52 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.down_proj_weight_quantizers.53 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.down_proj_weight_quantizers.54 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.15e-01 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.down_proj_weight_quantizers.55 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.54e-02 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.down_proj_weight_quantizers.56 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.down_proj_weight_quantizers.57 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.down_proj_weight_quantizers.58 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.50e-02 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.down_proj_weight_quantizers.59 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.59e-02 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.down_proj_weight_quantizers.60 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.down_proj_weight_quantizers.61 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.down_proj_weight_quantizers.62 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.down_proj_weight_quantizers.63 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.down_proj_weight_quantizers.64 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.down_proj_weight_quantizers.65 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.down_proj_weight_quantizers.66 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.down_proj_weight_quantizers.67 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.down_proj_weight_quantizers.68 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.down_proj_weight_quantizers.69 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.down_proj_weight_quantizers.70 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.down_proj_weight_quantizers.71 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.46e-01 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.down_proj_weight_quantizers.72 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.down_proj_weight_quantizers.73 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.down_proj_weight_quantizers.74 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.down_proj_weight_quantizers.75 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.down_proj_weight_quantizers.76 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.down_proj_weight_quantizers.77 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.down_proj_weight_quantizers.78 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.44e-01 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.down_proj_weight_quantizers.79 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.down_proj_weight_quantizers.80 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.down_proj_weight_quantizers.81 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.down_proj_weight_quantizers.82 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.24e-01 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.down_proj_weight_quantizers.83 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.47e-01 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.down_proj_weight_quantizers.84 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.01e-01 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.down_proj_weight_quantizers.85 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.55e-01 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.down_proj_weight_quantizers.86 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.down_proj_weight_quantizers.87 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.down_proj_weight_quantizers.88 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.down_proj_weight_quantizers.89 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.down_proj_weight_quantizers.90 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.29e-01 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.down_proj_weight_quantizers.91 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.down_proj_weight_quantizers.92 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.down_proj_weight_quantizers.93 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.down_proj_weight_quantizers.94 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.59e-02 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.down_proj_weight_quantizers.95 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.down_proj_weight_quantizers.96 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.56e-01 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.down_proj_weight_quantizers.97 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.50e-02 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.down_proj_weight_quantizers.98 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.down_proj_weight_quantizers.99 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.69e-01 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.down_proj_weight_quantizers.100 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.down_proj_weight_quantizers.101 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.down_proj_weight_quantizers.102 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.03e-01 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.down_proj_weight_quantizers.103 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.64e-02 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.down_proj_weight_quantizers.104 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.down_proj_weight_quantizers.105 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.down_proj_weight_quantizers.106 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.down_proj_weight_quantizers.107 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.40e-02 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.down_proj_weight_quantizers.108 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.down_proj_weight_quantizers.109 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.64e-02 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.down_proj_weight_quantizers.110 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.down_proj_weight_quantizers.111 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.down_proj_weight_quantizers.112 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.50e-02 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.down_proj_weight_quantizers.113 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.down_proj_weight_quantizers.114 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.down_proj_weight_quantizers.115 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.down_proj_weight_quantizers.116 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.down_proj_weight_quantizers.117 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.45e-02 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.down_proj_weight_quantizers.118 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.down_proj_weight_quantizers.119 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.45e-02 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.down_proj_weight_quantizers.120 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.55e-01 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.down_proj_weight_quantizers.121 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.21e-01 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.down_proj_weight_quantizers.122 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.down_proj_weight_quantizers.123 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.down_proj_weight_quantizers.124 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.down_proj_weight_quantizers.125 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.down_proj_weight_quantizers.126 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.down_proj_weight_quantizers.127 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.down_proj_weight_quantizers.128 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.down_proj_weight_quantizers.129 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.40e-02 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.down_proj_weight_quantizers.130 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.down_proj_weight_quantizers.131 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.54e-02 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.down_proj_weight_quantizers.132 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.down_proj_weight_quantizers.133 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.16e-01 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.down_proj_weight_quantizers.134 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.34e-01 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.down_proj_weight_quantizers.135 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.80e-01 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.down_proj_weight_quantizers.136 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.down_proj_weight_quantizers.137 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.down_proj_weight_quantizers.138 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.down_proj_weight_quantizers.139 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.16e-01 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.down_proj_weight_quantizers.140 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.26e-01 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.down_proj_weight_quantizers.141 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.down_proj_weight_quantizers.142 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.40e-02 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.down_proj_weight_quantizers.143 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.50e-02 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.down_proj_weight_quantizers.144 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.down_proj_weight_quantizers.145 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.26e-01 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.down_proj_weight_quantizers.146 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.52e-01 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.down_proj_weight_quantizers.147 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.down_proj_weight_quantizers.148 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.down_proj_weight_quantizers.149 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.down_proj_weight_quantizers.150 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.down_proj_weight_quantizers.151 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.down_proj_weight_quantizers.152 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.down_proj_weight_quantizers.153 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.down_proj_weight_quantizers.154 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.down_proj_weight_quantizers.155 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.down_proj_weight_quantizers.156 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.down_proj_weight_quantizers.157 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.down_proj_weight_quantizers.158 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.12e-01 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.down_proj_weight_quantizers.159 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.64e-02 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.down_proj_weight_quantizers.160 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.down_proj_weight_quantizers.161 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.down_proj_weight_quantizers.162 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.41e-01 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.down_proj_weight_quantizers.163 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.down_proj_weight_quantizers.164 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.down_proj_weight_quantizers.165 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.down_proj_weight_quantizers.166 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.down_proj_weight_quantizers.167 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.down_proj_weight_quantizers.168 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.down_proj_weight_quantizers.169 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.down_proj_weight_quantizers.170 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.73e-01 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.down_proj_weight_quantizers.171 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.down_proj_weight_quantizers.172 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.43e-01 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.down_proj_weight_quantizers.173 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.down_proj_weight_quantizers.174 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.down_proj_weight_quantizers.175 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.down_proj_weight_quantizers.176 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.33e-01 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.down_proj_weight_quantizers.177 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.66e-01 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.down_proj_weight_quantizers.178 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.64e-02 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.down_proj_weight_quantizers.179 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.down_proj_weight_quantizers.180 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.16e-01 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.down_proj_weight_quantizers.181 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.84e-01 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.down_proj_weight_quantizers.182 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.down_proj_weight_quantizers.183 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.64e-02 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.down_proj_weight_quantizers.184 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.down_proj_weight_quantizers.185 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.down_proj_weight_quantizers.186 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.13e-01 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.down_proj_weight_quantizers.187 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.down_proj_weight_quantizers.188 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.down_proj_weight_quantizers.189 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.66e-01 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.down_proj_weight_quantizers.190 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.down_proj_weight_quantizers.191 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.down_proj_weight_quantizers.192 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.down_proj_weight_quantizers.193 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.45e-02 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.down_proj_weight_quantizers.194 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.down_proj_weight_quantizers.195 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.64e-02 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.down_proj_weight_quantizers.196 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.down_proj_weight_quantizers.197 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.down_proj_weight_quantizers.198 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.down_proj_weight_quantizers.199 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.down_proj_weight_quantizers.200 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.37e-01 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.down_proj_weight_quantizers.201 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.down_proj_weight_quantizers.202 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.down_proj_weight_quantizers.203 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.down_proj_weight_quantizers.204 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.down_proj_weight_quantizers.205 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.59e-02 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.down_proj_weight_quantizers.206 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.down_proj_weight_quantizers.207 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.down_proj_weight_quantizers.208 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.down_proj_weight_quantizers.209 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.down_proj_weight_quantizers.210 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.down_proj_weight_quantizers.211 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.down_proj_weight_quantizers.212 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.down_proj_weight_quantizers.213 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.down_proj_weight_quantizers.214 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.57e-01 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.down_proj_weight_quantizers.215 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.down_proj_weight_quantizers.216 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.down_proj_weight_quantizers.217 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.down_proj_weight_quantizers.218 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.down_proj_weight_quantizers.219 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.down_proj_weight_quantizers.220 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.89e-01 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.down_proj_weight_quantizers.221 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.down_proj_weight_quantizers.222 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.down_proj_weight_quantizers.223 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.down_proj_weight_quantizers.224 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.down_proj_weight_quantizers.225 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.down_proj_weight_quantizers.226 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.down_proj_weight_quantizers.227 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.down_proj_weight_quantizers.228 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.down_proj_weight_quantizers.229 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.down_proj_weight_quantizers.230 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.down_proj_weight_quantizers.231 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.down_proj_weight_quantizers.232 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.41e-01 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.down_proj_weight_quantizers.233 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.down_proj_weight_quantizers.234 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.down_proj_weight_quantizers.235 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.down_proj_weight_quantizers.236 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.40e-02 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.down_proj_weight_quantizers.237 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.down_proj_weight_quantizers.238 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.down_proj_weight_quantizers.239 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.down_proj_weight_quantizers.240 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.20e-01 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.down_proj_weight_quantizers.241 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.45e-02 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.down_proj_weight_quantizers.242 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.down_proj_weight_quantizers.243 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.down_proj_weight_quantizers.244 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.down_proj_weight_quantizers.245 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.down_proj_weight_quantizers.246 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.57e-01 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.down_proj_weight_quantizers.247 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.down_proj_weight_quantizers.248 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.59e-02 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.down_proj_weight_quantizers.249 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.down_proj_weight_quantizers.250 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.down_proj_weight_quantizers.251 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.down_proj_weight_quantizers.252 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.down_proj_weight_quantizers.253 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.down_proj_weight_quantizers.254 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.20e-02 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.down_proj_weight_quantizers.255 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.18e-01 calibrator=MaxCalibrator quant) +model.layers.34.mlp.shared_experts.gate_proj.input_quantizer TensorQuantizer(disabled) +model.layers.34.mlp.shared_experts.gate_proj.output_quantizer TensorQuantizer(disabled) +model.layers.34.mlp.shared_experts.gate_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.34.mlp.shared_experts.up_proj.input_quantizer TensorQuantizer(disabled) +model.layers.34.mlp.shared_experts.up_proj.output_quantizer TensorQuantizer(disabled) +model.layers.34.mlp.shared_experts.up_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.34.mlp.shared_experts.down_proj.input_quantizer TensorQuantizer(disabled) +model.layers.34.mlp.shared_experts.down_proj.output_quantizer TensorQuantizer(disabled) +model.layers.34.mlp.shared_experts.down_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.35.self_attn.q_a_proj.input_quantizer TensorQuantizer(disabled) +model.layers.35.self_attn.q_a_proj.output_quantizer TensorQuantizer(disabled) +model.layers.35.self_attn.q_a_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.35.self_attn.q_b_proj.input_quantizer TensorQuantizer(disabled) +model.layers.35.self_attn.q_b_proj.output_quantizer TensorQuantizer(disabled) +model.layers.35.self_attn.q_b_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.35.self_attn.kv_a_proj_with_mqa.input_quantizer TensorQuantizer(disabled) +model.layers.35.self_attn.kv_a_proj_with_mqa.output_quantizer TensorQuantizer(disabled) +model.layers.35.self_attn.kv_a_proj_with_mqa.weight_quantizer TensorQuantizer(disabled) +model.layers.35.self_attn.kv_b_proj.input_quantizer TensorQuantizer(disabled) +model.layers.35.self_attn.kv_b_proj.output_quantizer TensorQuantizer(disabled) +model.layers.35.self_attn.kv_b_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.35.self_attn.o_proj.input_quantizer TensorQuantizer(disabled) +model.layers.35.self_attn.o_proj.output_quantizer TensorQuantizer(disabled) +model.layers.35.self_attn.o_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.35.self_attn.q_bmm_quantizer TensorQuantizer(disabled) +model.layers.35.self_attn.k_bmm_quantizer TensorQuantizer((4, 3) bit fake per-tensor amax=4.48e+02(const) calibrator=MaxCalibrator quant) +model.layers.35.self_attn.v_bmm_quantizer TensorQuantizer((4, 3) bit fake per-tensor amax=4.48e+02(const) calibrator=MaxCalibrator quant) +model.layers.35.self_attn.p_bmm_quantizer TensorQuantizer(disabled) +model.layers.35.mlp.experts.gate_up_proj_input_quantizer TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.67e+00 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.gate_up_proj_weight_quantizers.0 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.gate_up_proj_weight_quantizers.1 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.66e-01 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.gate_up_proj_weight_quantizers.2 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.25e-01 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.gate_up_proj_weight_quantizers.3 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.gate_up_proj_weight_quantizers.4 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.gate_up_proj_weight_quantizers.5 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.gate_up_proj_weight_quantizers.6 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.65e-01 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.gate_up_proj_weight_quantizers.7 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.gate_up_proj_weight_quantizers.8 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.gate_up_proj_weight_quantizers.9 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.23e-01 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.gate_up_proj_weight_quantizers.10 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.gate_up_proj_weight_quantizers.11 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=6.84e-01 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.gate_up_proj_weight_quantizers.12 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.gate_up_proj_weight_quantizers.13 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.gate_up_proj_weight_quantizers.14 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.gate_up_proj_weight_quantizers.15 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.gate_up_proj_weight_quantizers.16 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.gate_up_proj_weight_quantizers.17 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.17e-01 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.gate_up_proj_weight_quantizers.18 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.gate_up_proj_weight_quantizers.19 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.gate_up_proj_weight_quantizers.20 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.gate_up_proj_weight_quantizers.21 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.gate_up_proj_weight_quantizers.22 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.24e-01 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.gate_up_proj_weight_quantizers.23 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.28e-01 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.gate_up_proj_weight_quantizers.24 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.49e-01 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.gate_up_proj_weight_quantizers.25 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.gate_up_proj_weight_quantizers.26 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.gate_up_proj_weight_quantizers.27 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.44e-01 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.gate_up_proj_weight_quantizers.28 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.gate_up_proj_weight_quantizers.29 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.gate_up_proj_weight_quantizers.30 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.gate_up_proj_weight_quantizers.31 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.gate_up_proj_weight_quantizers.32 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.gate_up_proj_weight_quantizers.33 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.gate_up_proj_weight_quantizers.34 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.81e-01 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.gate_up_proj_weight_quantizers.35 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.gate_up_proj_weight_quantizers.36 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.gate_up_proj_weight_quantizers.37 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.30e-01 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.gate_up_proj_weight_quantizers.38 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.gate_up_proj_weight_quantizers.39 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.gate_up_proj_weight_quantizers.40 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.36e-01 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.gate_up_proj_weight_quantizers.41 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.gate_up_proj_weight_quantizers.42 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.gate_up_proj_weight_quantizers.43 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.gate_up_proj_weight_quantizers.44 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.gate_up_proj_weight_quantizers.45 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.13e-01 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.gate_up_proj_weight_quantizers.46 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.16e-01 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.gate_up_proj_weight_quantizers.47 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.gate_up_proj_weight_quantizers.48 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.36e-01 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.gate_up_proj_weight_quantizers.49 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.13e-01 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.gate_up_proj_weight_quantizers.50 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.gate_up_proj_weight_quantizers.51 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.gate_up_proj_weight_quantizers.52 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.gate_up_proj_weight_quantizers.53 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.gate_up_proj_weight_quantizers.54 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.gate_up_proj_weight_quantizers.55 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.gate_up_proj_weight_quantizers.56 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.gate_up_proj_weight_quantizers.57 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.gate_up_proj_weight_quantizers.58 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.27e-01 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.gate_up_proj_weight_quantizers.59 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.54e-01 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.gate_up_proj_weight_quantizers.60 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.gate_up_proj_weight_quantizers.61 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.gate_up_proj_weight_quantizers.62 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.gate_up_proj_weight_quantizers.63 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.34e-01 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.gate_up_proj_weight_quantizers.64 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.gate_up_proj_weight_quantizers.65 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.27e-01 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.gate_up_proj_weight_quantizers.66 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.gate_up_proj_weight_quantizers.67 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.gate_up_proj_weight_quantizers.68 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.gate_up_proj_weight_quantizers.69 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.24e-01 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.gate_up_proj_weight_quantizers.70 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.gate_up_proj_weight_quantizers.71 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.24e-01 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.gate_up_proj_weight_quantizers.72 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.29e-01 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.gate_up_proj_weight_quantizers.73 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.gate_up_proj_weight_quantizers.74 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.gate_up_proj_weight_quantizers.75 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.gate_up_proj_weight_quantizers.76 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.16e-01 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.gate_up_proj_weight_quantizers.77 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.gate_up_proj_weight_quantizers.78 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.82e-01 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.gate_up_proj_weight_quantizers.79 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.13e-01 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.gate_up_proj_weight_quantizers.80 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.gate_up_proj_weight_quantizers.81 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.gate_up_proj_weight_quantizers.82 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.gate_up_proj_weight_quantizers.83 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.gate_up_proj_weight_quantizers.84 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.65e-01 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.gate_up_proj_weight_quantizers.85 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.77e-01 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.gate_up_proj_weight_quantizers.86 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.gate_up_proj_weight_quantizers.87 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.gate_up_proj_weight_quantizers.88 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.gate_up_proj_weight_quantizers.89 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.gate_up_proj_weight_quantizers.90 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.gate_up_proj_weight_quantizers.91 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.gate_up_proj_weight_quantizers.92 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.gate_up_proj_weight_quantizers.93 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.gate_up_proj_weight_quantizers.94 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.13e-01 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.gate_up_proj_weight_quantizers.95 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.06e-01 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.gate_up_proj_weight_quantizers.96 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.gate_up_proj_weight_quantizers.97 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.25e-01 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.gate_up_proj_weight_quantizers.98 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.gate_up_proj_weight_quantizers.99 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.gate_up_proj_weight_quantizers.100 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.19e-01 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.gate_up_proj_weight_quantizers.101 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.gate_up_proj_weight_quantizers.102 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.gate_up_proj_weight_quantizers.103 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.42e-01 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.gate_up_proj_weight_quantizers.104 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.61e-01 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.gate_up_proj_weight_quantizers.105 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.gate_up_proj_weight_quantizers.106 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.gate_up_proj_weight_quantizers.107 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.gate_up_proj_weight_quantizers.108 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.gate_up_proj_weight_quantizers.109 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.51e-01 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.gate_up_proj_weight_quantizers.110 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.40e-01 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.gate_up_proj_weight_quantizers.111 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.gate_up_proj_weight_quantizers.112 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.18e-01 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.gate_up_proj_weight_quantizers.113 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.gate_up_proj_weight_quantizers.114 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.gate_up_proj_weight_quantizers.115 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.gate_up_proj_weight_quantizers.116 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.gate_up_proj_weight_quantizers.117 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.16e-01 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.gate_up_proj_weight_quantizers.118 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.gate_up_proj_weight_quantizers.119 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.39e-01 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.gate_up_proj_weight_quantizers.120 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.34e-01 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.gate_up_proj_weight_quantizers.121 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.gate_up_proj_weight_quantizers.122 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.29e-01 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.gate_up_proj_weight_quantizers.123 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.63e-01 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.gate_up_proj_weight_quantizers.124 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.gate_up_proj_weight_quantizers.125 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.20e-01 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.gate_up_proj_weight_quantizers.126 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.64e-02 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.gate_up_proj_weight_quantizers.127 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=5.12e-01 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.gate_up_proj_weight_quantizers.128 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.95e-01 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.gate_up_proj_weight_quantizers.129 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.gate_up_proj_weight_quantizers.130 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.gate_up_proj_weight_quantizers.131 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.gate_up_proj_weight_quantizers.132 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.gate_up_proj_weight_quantizers.133 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.gate_up_proj_weight_quantizers.134 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.gate_up_proj_weight_quantizers.135 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.gate_up_proj_weight_quantizers.136 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.gate_up_proj_weight_quantizers.137 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.gate_up_proj_weight_quantizers.138 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.gate_up_proj_weight_quantizers.139 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.16e-01 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.gate_up_proj_weight_quantizers.140 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.gate_up_proj_weight_quantizers.141 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.27e-01 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.gate_up_proj_weight_quantizers.142 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.gate_up_proj_weight_quantizers.143 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.15e-01 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.gate_up_proj_weight_quantizers.144 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.gate_up_proj_weight_quantizers.145 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.gate_up_proj_weight_quantizers.146 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.gate_up_proj_weight_quantizers.147 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.87e-01 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.gate_up_proj_weight_quantizers.148 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.29e-01 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.gate_up_proj_weight_quantizers.149 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.20e-01 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.gate_up_proj_weight_quantizers.150 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.gate_up_proj_weight_quantizers.151 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.23e-01 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.gate_up_proj_weight_quantizers.152 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=4.08e-01 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.gate_up_proj_weight_quantizers.153 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.gate_up_proj_weight_quantizers.154 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.gate_up_proj_weight_quantizers.155 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.gate_up_proj_weight_quantizers.156 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.gate_up_proj_weight_quantizers.157 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.23e-01 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.gate_up_proj_weight_quantizers.158 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.91e-01 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.gate_up_proj_weight_quantizers.159 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.gate_up_proj_weight_quantizers.160 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.55e-01 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.gate_up_proj_weight_quantizers.161 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.27e-01 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.gate_up_proj_weight_quantizers.162 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.15e-01 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.gate_up_proj_weight_quantizers.163 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.24e-01 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.gate_up_proj_weight_quantizers.164 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.gate_up_proj_weight_quantizers.165 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.46e-01 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.gate_up_proj_weight_quantizers.166 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.15e-01 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.gate_up_proj_weight_quantizers.167 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.47e-01 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.gate_up_proj_weight_quantizers.168 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.gate_up_proj_weight_quantizers.169 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.40e-01 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.gate_up_proj_weight_quantizers.170 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.gate_up_proj_weight_quantizers.171 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.59e-02 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.gate_up_proj_weight_quantizers.172 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.gate_up_proj_weight_quantizers.173 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.gate_up_proj_weight_quantizers.174 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.gate_up_proj_weight_quantizers.175 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.gate_up_proj_weight_quantizers.176 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.gate_up_proj_weight_quantizers.177 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.64e-01 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.gate_up_proj_weight_quantizers.178 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.gate_up_proj_weight_quantizers.179 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.gate_up_proj_weight_quantizers.180 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.97e-01 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.gate_up_proj_weight_quantizers.181 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=4.10e-01 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.gate_up_proj_weight_quantizers.182 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.gate_up_proj_weight_quantizers.183 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.21e-01 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.gate_up_proj_weight_quantizers.184 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.gate_up_proj_weight_quantizers.185 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.gate_up_proj_weight_quantizers.186 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.gate_up_proj_weight_quantizers.187 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.gate_up_proj_weight_quantizers.188 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.gate_up_proj_weight_quantizers.189 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.38e-01 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.gate_up_proj_weight_quantizers.190 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.gate_up_proj_weight_quantizers.191 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.gate_up_proj_weight_quantizers.192 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.54e-01 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.gate_up_proj_weight_quantizers.193 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.30e-01 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.gate_up_proj_weight_quantizers.194 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.gate_up_proj_weight_quantizers.195 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.gate_up_proj_weight_quantizers.196 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.gate_up_proj_weight_quantizers.197 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.gate_up_proj_weight_quantizers.198 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.gate_up_proj_weight_quantizers.199 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.gate_up_proj_weight_quantizers.200 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.70e-01 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.gate_up_proj_weight_quantizers.201 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.gate_up_proj_weight_quantizers.202 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.gate_up_proj_weight_quantizers.203 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.gate_up_proj_weight_quantizers.204 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.54e-01 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.gate_up_proj_weight_quantizers.205 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.gate_up_proj_weight_quantizers.206 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.gate_up_proj_weight_quantizers.207 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.32e-01 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.gate_up_proj_weight_quantizers.208 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=5.20e-01 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.gate_up_proj_weight_quantizers.209 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.23e-01 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.gate_up_proj_weight_quantizers.210 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.gate_up_proj_weight_quantizers.211 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.27e-01 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.gate_up_proj_weight_quantizers.212 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.80e-01 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.gate_up_proj_weight_quantizers.213 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.gate_up_proj_weight_quantizers.214 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.gate_up_proj_weight_quantizers.215 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.gate_up_proj_weight_quantizers.216 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.gate_up_proj_weight_quantizers.217 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.66e-01 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.gate_up_proj_weight_quantizers.218 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.14e-01 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.gate_up_proj_weight_quantizers.219 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.gate_up_proj_weight_quantizers.220 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.gate_up_proj_weight_quantizers.221 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.gate_up_proj_weight_quantizers.222 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.35e-01 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.gate_up_proj_weight_quantizers.223 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.14e-01 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.gate_up_proj_weight_quantizers.224 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.gate_up_proj_weight_quantizers.225 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.51e-01 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.gate_up_proj_weight_quantizers.226 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.gate_up_proj_weight_quantizers.227 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.32e-01 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.gate_up_proj_weight_quantizers.228 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.gate_up_proj_weight_quantizers.229 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.gate_up_proj_weight_quantizers.230 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.71e-01 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.gate_up_proj_weight_quantizers.231 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.gate_up_proj_weight_quantizers.232 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.14e-01 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.gate_up_proj_weight_quantizers.233 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.gate_up_proj_weight_quantizers.234 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.gate_up_proj_weight_quantizers.235 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.gate_up_proj_weight_quantizers.236 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.gate_up_proj_weight_quantizers.237 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.gate_up_proj_weight_quantizers.238 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.gate_up_proj_weight_quantizers.239 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.gate_up_proj_weight_quantizers.240 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.41e-01 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.gate_up_proj_weight_quantizers.241 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.gate_up_proj_weight_quantizers.242 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.06e-01 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.gate_up_proj_weight_quantizers.243 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.gate_up_proj_weight_quantizers.244 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.71e-01 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.gate_up_proj_weight_quantizers.245 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.gate_up_proj_weight_quantizers.246 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.18e-01 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.gate_up_proj_weight_quantizers.247 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.54e-01 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.gate_up_proj_weight_quantizers.248 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.gate_up_proj_weight_quantizers.249 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.gate_up_proj_weight_quantizers.250 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.14e-01 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.gate_up_proj_weight_quantizers.251 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.gate_up_proj_weight_quantizers.252 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.01e-01 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.gate_up_proj_weight_quantizers.253 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.gate_up_proj_weight_quantizers.254 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.71e-01 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.gate_up_proj_weight_quantizers.255 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.down_proj_input_quantizer TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.44e+01 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.down_proj_weight_quantizers.0 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.33e-01 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.down_proj_weight_quantizers.1 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.down_proj_weight_quantizers.2 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.down_proj_weight_quantizers.3 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.down_proj_weight_quantizers.4 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.down_proj_weight_quantizers.5 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.down_proj_weight_quantizers.6 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.88e-01 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.down_proj_weight_quantizers.7 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.17e-01 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.down_proj_weight_quantizers.8 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.down_proj_weight_quantizers.9 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.down_proj_weight_quantizers.10 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.down_proj_weight_quantizers.11 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.19e-01 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.down_proj_weight_quantizers.12 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.54e-02 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.down_proj_weight_quantizers.13 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.down_proj_weight_quantizers.14 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.22e-01 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.down_proj_weight_quantizers.15 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.down_proj_weight_quantizers.16 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.down_proj_weight_quantizers.17 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.down_proj_weight_quantizers.18 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.down_proj_weight_quantizers.19 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.down_proj_weight_quantizers.20 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.down_proj_weight_quantizers.21 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.down_proj_weight_quantizers.22 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.down_proj_weight_quantizers.23 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.down_proj_weight_quantizers.24 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.56e-01 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.down_proj_weight_quantizers.25 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.down_proj_weight_quantizers.26 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.down_proj_weight_quantizers.27 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.20e-02 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.down_proj_weight_quantizers.28 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.down_proj_weight_quantizers.29 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.down_proj_weight_quantizers.30 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.down_proj_weight_quantizers.31 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.down_proj_weight_quantizers.32 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.40e-02 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.down_proj_weight_quantizers.33 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.down_proj_weight_quantizers.34 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.77e-01 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.down_proj_weight_quantizers.35 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.down_proj_weight_quantizers.36 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.54e-02 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.down_proj_weight_quantizers.37 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.down_proj_weight_quantizers.38 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.down_proj_weight_quantizers.39 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.down_proj_weight_quantizers.40 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.47e-01 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.down_proj_weight_quantizers.41 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.down_proj_weight_quantizers.42 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.down_proj_weight_quantizers.43 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.down_proj_weight_quantizers.44 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.59e-02 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.down_proj_weight_quantizers.45 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.down_proj_weight_quantizers.46 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.down_proj_weight_quantizers.47 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.down_proj_weight_quantizers.48 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.54e-02 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.down_proj_weight_quantizers.49 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.01e-01 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.down_proj_weight_quantizers.50 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.59e-02 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.down_proj_weight_quantizers.51 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.59e-02 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.down_proj_weight_quantizers.52 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.down_proj_weight_quantizers.53 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.down_proj_weight_quantizers.54 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.54e-02 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.down_proj_weight_quantizers.55 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.down_proj_weight_quantizers.56 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.57e-01 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.down_proj_weight_quantizers.57 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.down_proj_weight_quantizers.58 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.down_proj_weight_quantizers.59 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.down_proj_weight_quantizers.60 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.down_proj_weight_quantizers.61 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.down_proj_weight_quantizers.62 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.down_proj_weight_quantizers.63 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.down_proj_weight_quantizers.64 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.down_proj_weight_quantizers.65 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.69e-01 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.down_proj_weight_quantizers.66 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.down_proj_weight_quantizers.67 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.down_proj_weight_quantizers.68 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.down_proj_weight_quantizers.69 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.down_proj_weight_quantizers.70 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.down_proj_weight_quantizers.71 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.45e-01 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.down_proj_weight_quantizers.72 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.72e-01 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.down_proj_weight_quantizers.73 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.down_proj_weight_quantizers.74 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.23e-01 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.down_proj_weight_quantizers.75 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.down_proj_weight_quantizers.76 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.down_proj_weight_quantizers.77 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.down_proj_weight_quantizers.78 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.14e-01 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.down_proj_weight_quantizers.79 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.down_proj_weight_quantizers.80 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.down_proj_weight_quantizers.81 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.down_proj_weight_quantizers.82 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.down_proj_weight_quantizers.83 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.54e-02 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.down_proj_weight_quantizers.84 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.32e-01 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.down_proj_weight_quantizers.85 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.83e-01 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.down_proj_weight_quantizers.86 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.down_proj_weight_quantizers.87 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.59e-02 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.down_proj_weight_quantizers.88 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.down_proj_weight_quantizers.89 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.99e-01 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.down_proj_weight_quantizers.90 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.down_proj_weight_quantizers.91 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.down_proj_weight_quantizers.92 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.down_proj_weight_quantizers.93 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.down_proj_weight_quantizers.94 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.21e-01 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.down_proj_weight_quantizers.95 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.29e-01 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.down_proj_weight_quantizers.96 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.down_proj_weight_quantizers.97 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.down_proj_weight_quantizers.98 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.down_proj_weight_quantizers.99 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.down_proj_weight_quantizers.100 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.down_proj_weight_quantizers.101 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.40e-02 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.down_proj_weight_quantizers.102 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.down_proj_weight_quantizers.103 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.32e-01 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.down_proj_weight_quantizers.104 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.45e-02 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.down_proj_weight_quantizers.105 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.down_proj_weight_quantizers.106 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.down_proj_weight_quantizers.107 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.21e-01 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.down_proj_weight_quantizers.108 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.down_proj_weight_quantizers.109 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.60e-01 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.down_proj_weight_quantizers.110 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.down_proj_weight_quantizers.111 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.down_proj_weight_quantizers.112 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.down_proj_weight_quantizers.113 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.down_proj_weight_quantizers.114 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.40e-02 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.down_proj_weight_quantizers.115 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.down_proj_weight_quantizers.116 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.39e-01 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.down_proj_weight_quantizers.117 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.28e-01 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.down_proj_weight_quantizers.118 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.down_proj_weight_quantizers.119 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.64e-02 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.down_proj_weight_quantizers.120 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.down_proj_weight_quantizers.121 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.down_proj_weight_quantizers.122 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.29e-01 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.down_proj_weight_quantizers.123 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.24e-01 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.down_proj_weight_quantizers.124 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.down_proj_weight_quantizers.125 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.down_proj_weight_quantizers.126 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.down_proj_weight_quantizers.127 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.34e-01 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.down_proj_weight_quantizers.128 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.down_proj_weight_quantizers.129 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.down_proj_weight_quantizers.130 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.down_proj_weight_quantizers.131 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.50e-02 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.down_proj_weight_quantizers.132 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.down_proj_weight_quantizers.133 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.54e-02 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.down_proj_weight_quantizers.134 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.down_proj_weight_quantizers.135 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.down_proj_weight_quantizers.136 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.down_proj_weight_quantizers.137 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.down_proj_weight_quantizers.138 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.down_proj_weight_quantizers.139 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.26e-01 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.down_proj_weight_quantizers.140 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.down_proj_weight_quantizers.141 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.52e-01 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.down_proj_weight_quantizers.142 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.down_proj_weight_quantizers.143 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.down_proj_weight_quantizers.144 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.50e-02 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.down_proj_weight_quantizers.145 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.down_proj_weight_quantizers.146 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.down_proj_weight_quantizers.147 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.99e-01 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.down_proj_weight_quantizers.148 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.down_proj_weight_quantizers.149 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.down_proj_weight_quantizers.150 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.down_proj_weight_quantizers.151 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.down_proj_weight_quantizers.152 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.25e-01 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.down_proj_weight_quantizers.153 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.down_proj_weight_quantizers.154 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.down_proj_weight_quantizers.155 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.down_proj_weight_quantizers.156 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.92e-01 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.down_proj_weight_quantizers.157 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.down_proj_weight_quantizers.158 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.40e-01 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.down_proj_weight_quantizers.159 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.45e-02 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.down_proj_weight_quantizers.160 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.42e-01 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.down_proj_weight_quantizers.161 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.down_proj_weight_quantizers.162 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.64e-02 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.down_proj_weight_quantizers.163 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.down_proj_weight_quantizers.164 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.down_proj_weight_quantizers.165 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.down_proj_weight_quantizers.166 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.down_proj_weight_quantizers.167 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.81e-01 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.down_proj_weight_quantizers.168 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.down_proj_weight_quantizers.169 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.down_proj_weight_quantizers.170 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.down_proj_weight_quantizers.171 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.down_proj_weight_quantizers.172 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.down_proj_weight_quantizers.173 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.down_proj_weight_quantizers.174 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.down_proj_weight_quantizers.175 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.40e-02 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.down_proj_weight_quantizers.176 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.down_proj_weight_quantizers.177 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.71e-01 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.down_proj_weight_quantizers.178 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.59e-02 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.down_proj_weight_quantizers.179 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.down_proj_weight_quantizers.180 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.22e-01 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.down_proj_weight_quantizers.181 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.63e-01 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.down_proj_weight_quantizers.182 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.59e-02 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.down_proj_weight_quantizers.183 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.down_proj_weight_quantizers.184 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.down_proj_weight_quantizers.185 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.down_proj_weight_quantizers.186 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.64e-02 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.down_proj_weight_quantizers.187 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.down_proj_weight_quantizers.188 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.down_proj_weight_quantizers.189 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.56e-01 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.down_proj_weight_quantizers.190 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.down_proj_weight_quantizers.191 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.down_proj_weight_quantizers.192 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.62e-01 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.down_proj_weight_quantizers.193 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.down_proj_weight_quantizers.194 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.down_proj_weight_quantizers.195 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.down_proj_weight_quantizers.196 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.down_proj_weight_quantizers.197 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.down_proj_weight_quantizers.198 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.54e-02 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.down_proj_weight_quantizers.199 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.73e-01 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.down_proj_weight_quantizers.200 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.down_proj_weight_quantizers.201 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.59e-02 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.down_proj_weight_quantizers.202 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.down_proj_weight_quantizers.203 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.down_proj_weight_quantizers.204 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.down_proj_weight_quantizers.205 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.down_proj_weight_quantizers.206 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.down_proj_weight_quantizers.207 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.40e-01 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.down_proj_weight_quantizers.208 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.10e-01 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.down_proj_weight_quantizers.209 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.29e-01 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.down_proj_weight_quantizers.210 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.54e-02 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.down_proj_weight_quantizers.211 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.35e-02 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.down_proj_weight_quantizers.212 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.40e-01 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.down_proj_weight_quantizers.213 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.down_proj_weight_quantizers.214 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.64e-02 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.down_proj_weight_quantizers.215 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.down_proj_weight_quantizers.216 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.down_proj_weight_quantizers.217 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.down_proj_weight_quantizers.218 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.down_proj_weight_quantizers.219 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.54e-02 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.down_proj_weight_quantizers.220 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.down_proj_weight_quantizers.221 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.down_proj_weight_quantizers.222 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.36e-01 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.down_proj_weight_quantizers.223 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.down_proj_weight_quantizers.224 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.down_proj_weight_quantizers.225 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.90e-01 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.down_proj_weight_quantizers.226 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.down_proj_weight_quantizers.227 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.64e-02 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.down_proj_weight_quantizers.228 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.down_proj_weight_quantizers.229 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.down_proj_weight_quantizers.230 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.00e-01 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.down_proj_weight_quantizers.231 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.down_proj_weight_quantizers.232 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.down_proj_weight_quantizers.233 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.down_proj_weight_quantizers.234 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.down_proj_weight_quantizers.235 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.down_proj_weight_quantizers.236 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.down_proj_weight_quantizers.237 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.45e-02 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.down_proj_weight_quantizers.238 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.down_proj_weight_quantizers.239 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.down_proj_weight_quantizers.240 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.down_proj_weight_quantizers.241 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.down_proj_weight_quantizers.242 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.51e-01 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.down_proj_weight_quantizers.243 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.59e-02 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.down_proj_weight_quantizers.244 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.down_proj_weight_quantizers.245 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.down_proj_weight_quantizers.246 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.down_proj_weight_quantizers.247 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.54e-02 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.down_proj_weight_quantizers.248 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.down_proj_weight_quantizers.249 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.down_proj_weight_quantizers.250 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.down_proj_weight_quantizers.251 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.95e-01 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.down_proj_weight_quantizers.252 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.down_proj_weight_quantizers.253 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.down_proj_weight_quantizers.254 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.27e-01 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.down_proj_weight_quantizers.255 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.35.mlp.shared_experts.gate_proj.input_quantizer TensorQuantizer(disabled) +model.layers.35.mlp.shared_experts.gate_proj.output_quantizer TensorQuantizer(disabled) +model.layers.35.mlp.shared_experts.gate_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.35.mlp.shared_experts.up_proj.input_quantizer TensorQuantizer(disabled) +model.layers.35.mlp.shared_experts.up_proj.output_quantizer TensorQuantizer(disabled) +model.layers.35.mlp.shared_experts.up_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.35.mlp.shared_experts.down_proj.input_quantizer TensorQuantizer(disabled) +model.layers.35.mlp.shared_experts.down_proj.output_quantizer TensorQuantizer(disabled) +model.layers.35.mlp.shared_experts.down_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.36.self_attn.q_a_proj.input_quantizer TensorQuantizer(disabled) +model.layers.36.self_attn.q_a_proj.output_quantizer TensorQuantizer(disabled) +model.layers.36.self_attn.q_a_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.36.self_attn.q_b_proj.input_quantizer TensorQuantizer(disabled) +model.layers.36.self_attn.q_b_proj.output_quantizer TensorQuantizer(disabled) +model.layers.36.self_attn.q_b_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.36.self_attn.kv_a_proj_with_mqa.input_quantizer TensorQuantizer(disabled) +model.layers.36.self_attn.kv_a_proj_with_mqa.output_quantizer TensorQuantizer(disabled) +model.layers.36.self_attn.kv_a_proj_with_mqa.weight_quantizer TensorQuantizer(disabled) +model.layers.36.self_attn.kv_b_proj.input_quantizer TensorQuantizer(disabled) +model.layers.36.self_attn.kv_b_proj.output_quantizer TensorQuantizer(disabled) +model.layers.36.self_attn.kv_b_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.36.self_attn.o_proj.input_quantizer TensorQuantizer(disabled) +model.layers.36.self_attn.o_proj.output_quantizer TensorQuantizer(disabled) +model.layers.36.self_attn.o_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.36.self_attn.q_bmm_quantizer TensorQuantizer(disabled) +model.layers.36.self_attn.k_bmm_quantizer TensorQuantizer((4, 3) bit fake per-tensor amax=4.48e+02(const) calibrator=MaxCalibrator quant) +model.layers.36.self_attn.v_bmm_quantizer TensorQuantizer((4, 3) bit fake per-tensor amax=4.48e+02(const) calibrator=MaxCalibrator quant) +model.layers.36.self_attn.p_bmm_quantizer TensorQuantizer(disabled) +model.layers.36.mlp.experts.gate_up_proj_input_quantizer TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.61e+00 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.gate_up_proj_weight_quantizers.0 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.13e-01 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.gate_up_proj_weight_quantizers.1 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.gate_up_proj_weight_quantizers.2 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.gate_up_proj_weight_quantizers.3 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.gate_up_proj_weight_quantizers.4 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.gate_up_proj_weight_quantizers.5 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.gate_up_proj_weight_quantizers.6 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.52e-01 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.gate_up_proj_weight_quantizers.7 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.77e-01 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.gate_up_proj_weight_quantizers.8 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.gate_up_proj_weight_quantizers.9 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.gate_up_proj_weight_quantizers.10 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.28e-01 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.gate_up_proj_weight_quantizers.11 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.gate_up_proj_weight_quantizers.12 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.gate_up_proj_weight_quantizers.13 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.gate_up_proj_weight_quantizers.14 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.gate_up_proj_weight_quantizers.15 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.40e-01 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.gate_up_proj_weight_quantizers.16 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.gate_up_proj_weight_quantizers.17 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.72e-01 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.gate_up_proj_weight_quantizers.18 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.gate_up_proj_weight_quantizers.19 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.gate_up_proj_weight_quantizers.20 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.gate_up_proj_weight_quantizers.21 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.gate_up_proj_weight_quantizers.22 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.gate_up_proj_weight_quantizers.23 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.14e-01 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.gate_up_proj_weight_quantizers.24 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.gate_up_proj_weight_quantizers.25 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.91e-01 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.gate_up_proj_weight_quantizers.26 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.28e-01 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.gate_up_proj_weight_quantizers.27 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.gate_up_proj_weight_quantizers.28 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.gate_up_proj_weight_quantizers.29 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.gate_up_proj_weight_quantizers.30 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.gate_up_proj_weight_quantizers.31 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.gate_up_proj_weight_quantizers.32 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.gate_up_proj_weight_quantizers.33 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.79e-01 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.gate_up_proj_weight_quantizers.34 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.gate_up_proj_weight_quantizers.35 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.18e-01 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.gate_up_proj_weight_quantizers.36 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.gate_up_proj_weight_quantizers.37 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.gate_up_proj_weight_quantizers.38 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.50e-01 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.gate_up_proj_weight_quantizers.39 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.gate_up_proj_weight_quantizers.40 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.gate_up_proj_weight_quantizers.41 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=4.80e-01 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.gate_up_proj_weight_quantizers.42 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.gate_up_proj_weight_quantizers.43 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.24e-01 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.gate_up_proj_weight_quantizers.44 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.gate_up_proj_weight_quantizers.45 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.gate_up_proj_weight_quantizers.46 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.gate_up_proj_weight_quantizers.47 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.36e-01 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.gate_up_proj_weight_quantizers.48 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.gate_up_proj_weight_quantizers.49 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.21e-01 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.gate_up_proj_weight_quantizers.50 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.gate_up_proj_weight_quantizers.51 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.96e-01 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.gate_up_proj_weight_quantizers.52 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.19e-01 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.gate_up_proj_weight_quantizers.53 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.24e-01 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.gate_up_proj_weight_quantizers.54 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.gate_up_proj_weight_quantizers.55 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.23e-01 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.gate_up_proj_weight_quantizers.56 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.gate_up_proj_weight_quantizers.57 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.19e-01 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.gate_up_proj_weight_quantizers.58 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.gate_up_proj_weight_quantizers.59 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.42e-01 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.gate_up_proj_weight_quantizers.60 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.gate_up_proj_weight_quantizers.61 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.gate_up_proj_weight_quantizers.62 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.gate_up_proj_weight_quantizers.63 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.00e-01 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.gate_up_proj_weight_quantizers.64 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.gate_up_proj_weight_quantizers.65 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.gate_up_proj_weight_quantizers.66 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.gate_up_proj_weight_quantizers.67 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.42e-01 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.gate_up_proj_weight_quantizers.68 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.gate_up_proj_weight_quantizers.69 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.31e-01 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.gate_up_proj_weight_quantizers.70 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.gate_up_proj_weight_quantizers.71 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.gate_up_proj_weight_quantizers.72 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.gate_up_proj_weight_quantizers.73 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.gate_up_proj_weight_quantizers.74 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.gate_up_proj_weight_quantizers.75 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.gate_up_proj_weight_quantizers.76 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.13e-01 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.gate_up_proj_weight_quantizers.77 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.gate_up_proj_weight_quantizers.78 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.53e-01 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.gate_up_proj_weight_quantizers.79 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.41e-01 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.gate_up_proj_weight_quantizers.80 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.gate_up_proj_weight_quantizers.81 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.gate_up_proj_weight_quantizers.82 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.gate_up_proj_weight_quantizers.83 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.gate_up_proj_weight_quantizers.84 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.gate_up_proj_weight_quantizers.85 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.gate_up_proj_weight_quantizers.86 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.gate_up_proj_weight_quantizers.87 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.gate_up_proj_weight_quantizers.88 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.34e-01 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.gate_up_proj_weight_quantizers.89 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.28e-01 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.gate_up_proj_weight_quantizers.90 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.gate_up_proj_weight_quantizers.91 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.12e-01 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.gate_up_proj_weight_quantizers.92 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.gate_up_proj_weight_quantizers.93 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.gate_up_proj_weight_quantizers.94 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.19e-01 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.gate_up_proj_weight_quantizers.95 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.gate_up_proj_weight_quantizers.96 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.gate_up_proj_weight_quantizers.97 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.gate_up_proj_weight_quantizers.98 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.gate_up_proj_weight_quantizers.99 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.gate_up_proj_weight_quantizers.100 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.62e-01 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.gate_up_proj_weight_quantizers.101 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.45e-01 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.gate_up_proj_weight_quantizers.102 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.gate_up_proj_weight_quantizers.103 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.gate_up_proj_weight_quantizers.104 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.gate_up_proj_weight_quantizers.105 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.gate_up_proj_weight_quantizers.106 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.gate_up_proj_weight_quantizers.107 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.gate_up_proj_weight_quantizers.108 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.44e-01 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.gate_up_proj_weight_quantizers.109 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.gate_up_proj_weight_quantizers.110 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.32e-01 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.gate_up_proj_weight_quantizers.111 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.gate_up_proj_weight_quantizers.112 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.70e-01 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.gate_up_proj_weight_quantizers.113 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=4.47e-01 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.gate_up_proj_weight_quantizers.114 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.14e-01 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.gate_up_proj_weight_quantizers.115 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.gate_up_proj_weight_quantizers.116 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.gate_up_proj_weight_quantizers.117 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.gate_up_proj_weight_quantizers.118 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.gate_up_proj_weight_quantizers.119 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.78e-01 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.gate_up_proj_weight_quantizers.120 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.34e-01 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.gate_up_proj_weight_quantizers.121 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.22e-01 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.gate_up_proj_weight_quantizers.122 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.gate_up_proj_weight_quantizers.123 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.18e-01 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.gate_up_proj_weight_quantizers.124 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.gate_up_proj_weight_quantizers.125 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.gate_up_proj_weight_quantizers.126 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.gate_up_proj_weight_quantizers.127 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.18e-01 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.gate_up_proj_weight_quantizers.128 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.gate_up_proj_weight_quantizers.129 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.gate_up_proj_weight_quantizers.130 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.gate_up_proj_weight_quantizers.131 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.gate_up_proj_weight_quantizers.132 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.gate_up_proj_weight_quantizers.133 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.35e-01 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.gate_up_proj_weight_quantizers.134 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.gate_up_proj_weight_quantizers.135 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.70e-01 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.gate_up_proj_weight_quantizers.136 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.gate_up_proj_weight_quantizers.137 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.gate_up_proj_weight_quantizers.138 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.16e-01 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.gate_up_proj_weight_quantizers.139 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.80e-01 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.gate_up_proj_weight_quantizers.140 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.gate_up_proj_weight_quantizers.141 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.gate_up_proj_weight_quantizers.142 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.72e-01 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.gate_up_proj_weight_quantizers.143 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.gate_up_proj_weight_quantizers.144 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.26e-01 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.gate_up_proj_weight_quantizers.145 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.gate_up_proj_weight_quantizers.146 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.gate_up_proj_weight_quantizers.147 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.15e-01 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.gate_up_proj_weight_quantizers.148 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.gate_up_proj_weight_quantizers.149 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.gate_up_proj_weight_quantizers.150 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.gate_up_proj_weight_quantizers.151 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.gate_up_proj_weight_quantizers.152 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=4.16e-01 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.gate_up_proj_weight_quantizers.153 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.gate_up_proj_weight_quantizers.154 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=4.53e-01 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.gate_up_proj_weight_quantizers.155 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.79e-01 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.gate_up_proj_weight_quantizers.156 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.17e-01 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.gate_up_proj_weight_quantizers.157 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.81e-01 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.gate_up_proj_weight_quantizers.158 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.gate_up_proj_weight_quantizers.159 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.gate_up_proj_weight_quantizers.160 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.gate_up_proj_weight_quantizers.161 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=4.59e-01 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.gate_up_proj_weight_quantizers.162 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.46e-01 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.gate_up_proj_weight_quantizers.163 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.75e-01 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.gate_up_proj_weight_quantizers.164 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.gate_up_proj_weight_quantizers.165 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.gate_up_proj_weight_quantizers.166 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.37e-01 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.gate_up_proj_weight_quantizers.167 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.gate_up_proj_weight_quantizers.168 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.gate_up_proj_weight_quantizers.169 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.gate_up_proj_weight_quantizers.170 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.gate_up_proj_weight_quantizers.171 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.95e-01 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.gate_up_proj_weight_quantizers.172 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.gate_up_proj_weight_quantizers.173 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.49e-01 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.gate_up_proj_weight_quantizers.174 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.gate_up_proj_weight_quantizers.175 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.gate_up_proj_weight_quantizers.176 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.gate_up_proj_weight_quantizers.177 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.91e-01 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.gate_up_proj_weight_quantizers.178 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.52e-01 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.gate_up_proj_weight_quantizers.179 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.73e-01 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.gate_up_proj_weight_quantizers.180 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.gate_up_proj_weight_quantizers.181 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.gate_up_proj_weight_quantizers.182 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.71e-01 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.gate_up_proj_weight_quantizers.183 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.59e-02 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.gate_up_proj_weight_quantizers.184 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.57e-01 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.gate_up_proj_weight_quantizers.185 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.gate_up_proj_weight_quantizers.186 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.92e-01 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.gate_up_proj_weight_quantizers.187 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.gate_up_proj_weight_quantizers.188 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.19e-01 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.gate_up_proj_weight_quantizers.189 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.gate_up_proj_weight_quantizers.190 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.gate_up_proj_weight_quantizers.191 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.gate_up_proj_weight_quantizers.192 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.gate_up_proj_weight_quantizers.193 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.gate_up_proj_weight_quantizers.194 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.28e-01 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.gate_up_proj_weight_quantizers.195 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.39e-01 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.gate_up_proj_weight_quantizers.196 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.gate_up_proj_weight_quantizers.197 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.34e-01 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.gate_up_proj_weight_quantizers.198 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.15e-01 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.gate_up_proj_weight_quantizers.199 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.gate_up_proj_weight_quantizers.200 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.90e-01 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.gate_up_proj_weight_quantizers.201 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.gate_up_proj_weight_quantizers.202 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.gate_up_proj_weight_quantizers.203 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.30e-01 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.gate_up_proj_weight_quantizers.204 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.gate_up_proj_weight_quantizers.205 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.gate_up_proj_weight_quantizers.206 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.gate_up_proj_weight_quantizers.207 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.gate_up_proj_weight_quantizers.208 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.gate_up_proj_weight_quantizers.209 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.gate_up_proj_weight_quantizers.210 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.21e-01 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.gate_up_proj_weight_quantizers.211 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.19e-01 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.gate_up_proj_weight_quantizers.212 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.gate_up_proj_weight_quantizers.213 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.30e-01 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.gate_up_proj_weight_quantizers.214 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.37e-01 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.gate_up_proj_weight_quantizers.215 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.gate_up_proj_weight_quantizers.216 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.gate_up_proj_weight_quantizers.217 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.43e-01 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.gate_up_proj_weight_quantizers.218 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.gate_up_proj_weight_quantizers.219 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.gate_up_proj_weight_quantizers.220 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.13e-01 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.gate_up_proj_weight_quantizers.221 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.02e-01 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.gate_up_proj_weight_quantizers.222 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.56e-01 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.gate_up_proj_weight_quantizers.223 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.16e-01 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.gate_up_proj_weight_quantizers.224 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.55e-01 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.gate_up_proj_weight_quantizers.225 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.gate_up_proj_weight_quantizers.226 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.gate_up_proj_weight_quantizers.227 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.gate_up_proj_weight_quantizers.228 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.27e-01 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.gate_up_proj_weight_quantizers.229 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.16e-01 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.gate_up_proj_weight_quantizers.230 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.gate_up_proj_weight_quantizers.231 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.gate_up_proj_weight_quantizers.232 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.39e-01 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.gate_up_proj_weight_quantizers.233 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.30e-01 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.gate_up_proj_weight_quantizers.234 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.39e-01 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.gate_up_proj_weight_quantizers.235 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=4.18e-01 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.gate_up_proj_weight_quantizers.236 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.gate_up_proj_weight_quantizers.237 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=4.22e-01 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.gate_up_proj_weight_quantizers.238 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.gate_up_proj_weight_quantizers.239 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.gate_up_proj_weight_quantizers.240 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.13e-01 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.gate_up_proj_weight_quantizers.241 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.18e-01 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.gate_up_proj_weight_quantizers.242 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.gate_up_proj_weight_quantizers.243 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.97e-01 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.gate_up_proj_weight_quantizers.244 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.41e-01 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.gate_up_proj_weight_quantizers.245 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.gate_up_proj_weight_quantizers.246 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.gate_up_proj_weight_quantizers.247 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.gate_up_proj_weight_quantizers.248 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.40e-01 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.gate_up_proj_weight_quantizers.249 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.48e-01 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.gate_up_proj_weight_quantizers.250 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.13e-01 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.gate_up_proj_weight_quantizers.251 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.gate_up_proj_weight_quantizers.252 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.gate_up_proj_weight_quantizers.253 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.gate_up_proj_weight_quantizers.254 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.89e-01 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.gate_up_proj_weight_quantizers.255 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.46e-01 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.down_proj_input_quantizer TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.30e+01 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.down_proj_weight_quantizers.0 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.down_proj_weight_quantizers.1 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.down_proj_weight_quantizers.2 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.down_proj_weight_quantizers.3 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.down_proj_weight_quantizers.4 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.down_proj_weight_quantizers.5 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.down_proj_weight_quantizers.6 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.down_proj_weight_quantizers.7 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.down_proj_weight_quantizers.8 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.down_proj_weight_quantizers.9 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.down_proj_weight_quantizers.10 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.down_proj_weight_quantizers.11 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.down_proj_weight_quantizers.12 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.down_proj_weight_quantizers.13 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.down_proj_weight_quantizers.14 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.down_proj_weight_quantizers.15 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.down_proj_weight_quantizers.16 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.down_proj_weight_quantizers.17 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.down_proj_weight_quantizers.18 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.down_proj_weight_quantizers.19 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.50e-02 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.down_proj_weight_quantizers.20 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.46e-01 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.down_proj_weight_quantizers.21 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.22e-01 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.down_proj_weight_quantizers.22 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.down_proj_weight_quantizers.23 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.down_proj_weight_quantizers.24 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.down_proj_weight_quantizers.25 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.down_proj_weight_quantizers.26 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.down_proj_weight_quantizers.27 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.down_proj_weight_quantizers.28 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.64e-02 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.down_proj_weight_quantizers.29 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.down_proj_weight_quantizers.30 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.down_proj_weight_quantizers.31 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.down_proj_weight_quantizers.32 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.down_proj_weight_quantizers.33 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.down_proj_weight_quantizers.34 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.down_proj_weight_quantizers.35 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.down_proj_weight_quantizers.36 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.down_proj_weight_quantizers.37 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.down_proj_weight_quantizers.38 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.94e-01 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.down_proj_weight_quantizers.39 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.down_proj_weight_quantizers.40 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.down_proj_weight_quantizers.41 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.down_proj_weight_quantizers.42 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.54e-02 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.down_proj_weight_quantizers.43 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.down_proj_weight_quantizers.44 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.44e-01 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.down_proj_weight_quantizers.45 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.down_proj_weight_quantizers.46 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.down_proj_weight_quantizers.47 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.down_proj_weight_quantizers.48 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.down_proj_weight_quantizers.49 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.down_proj_weight_quantizers.50 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.down_proj_weight_quantizers.51 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.down_proj_weight_quantizers.52 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.down_proj_weight_quantizers.53 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.93e-01 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.down_proj_weight_quantizers.54 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.23e-01 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.down_proj_weight_quantizers.55 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.05e-01 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.down_proj_weight_quantizers.56 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.down_proj_weight_quantizers.57 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.down_proj_weight_quantizers.58 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.35e-02 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.down_proj_weight_quantizers.59 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.down_proj_weight_quantizers.60 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.down_proj_weight_quantizers.61 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.down_proj_weight_quantizers.62 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.down_proj_weight_quantizers.63 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.51e-01 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.down_proj_weight_quantizers.64 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.down_proj_weight_quantizers.65 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.down_proj_weight_quantizers.66 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.59e-02 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.down_proj_weight_quantizers.67 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.down_proj_weight_quantizers.68 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.down_proj_weight_quantizers.69 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.down_proj_weight_quantizers.70 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.down_proj_weight_quantizers.71 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.64e-02 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.down_proj_weight_quantizers.72 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.down_proj_weight_quantizers.73 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.down_proj_weight_quantizers.74 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.down_proj_weight_quantizers.75 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.down_proj_weight_quantizers.76 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.down_proj_weight_quantizers.77 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.down_proj_weight_quantizers.78 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.down_proj_weight_quantizers.79 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.64e-02 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.down_proj_weight_quantizers.80 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.down_proj_weight_quantizers.81 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.64e-02 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.down_proj_weight_quantizers.82 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.24e-01 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.down_proj_weight_quantizers.83 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.down_proj_weight_quantizers.84 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.down_proj_weight_quantizers.85 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.down_proj_weight_quantizers.86 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.down_proj_weight_quantizers.87 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.45e-02 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.down_proj_weight_quantizers.88 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.down_proj_weight_quantizers.89 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.down_proj_weight_quantizers.90 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.down_proj_weight_quantizers.91 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.05e-01 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.down_proj_weight_quantizers.92 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.59e-02 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.down_proj_weight_quantizers.93 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.down_proj_weight_quantizers.94 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.down_proj_weight_quantizers.95 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.down_proj_weight_quantizers.96 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.down_proj_weight_quantizers.97 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.down_proj_weight_quantizers.98 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.down_proj_weight_quantizers.99 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.down_proj_weight_quantizers.100 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.32e-01 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.down_proj_weight_quantizers.101 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.down_proj_weight_quantizers.102 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.down_proj_weight_quantizers.103 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.down_proj_weight_quantizers.104 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.64e-02 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.down_proj_weight_quantizers.105 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.down_proj_weight_quantizers.106 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.down_proj_weight_quantizers.107 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.down_proj_weight_quantizers.108 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.81e-01 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.down_proj_weight_quantizers.109 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.down_proj_weight_quantizers.110 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.down_proj_weight_quantizers.111 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.down_proj_weight_quantizers.112 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.down_proj_weight_quantizers.113 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.91e-01 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.down_proj_weight_quantizers.114 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.down_proj_weight_quantizers.115 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.down_proj_weight_quantizers.116 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.down_proj_weight_quantizers.117 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.59e-02 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.down_proj_weight_quantizers.118 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.64e-02 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.down_proj_weight_quantizers.119 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.64e-02 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.down_proj_weight_quantizers.120 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.17e-01 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.down_proj_weight_quantizers.121 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.18e-01 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.down_proj_weight_quantizers.122 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.down_proj_weight_quantizers.123 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.82e-01 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.down_proj_weight_quantizers.124 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.50e-02 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.down_proj_weight_quantizers.125 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.59e-02 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.down_proj_weight_quantizers.126 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.down_proj_weight_quantizers.127 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.down_proj_weight_quantizers.128 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.down_proj_weight_quantizers.129 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.down_proj_weight_quantizers.130 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.down_proj_weight_quantizers.131 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.down_proj_weight_quantizers.132 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.down_proj_weight_quantizers.133 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.down_proj_weight_quantizers.134 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.down_proj_weight_quantizers.135 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.down_proj_weight_quantizers.136 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.down_proj_weight_quantizers.137 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.down_proj_weight_quantizers.138 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.16e-01 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.down_proj_weight_quantizers.139 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.down_proj_weight_quantizers.140 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.down_proj_weight_quantizers.141 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.59e-02 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.down_proj_weight_quantizers.142 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.40e-01 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.down_proj_weight_quantizers.143 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.down_proj_weight_quantizers.144 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.down_proj_weight_quantizers.145 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.50e-02 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.down_proj_weight_quantizers.146 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.down_proj_weight_quantizers.147 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.down_proj_weight_quantizers.148 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.down_proj_weight_quantizers.149 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.down_proj_weight_quantizers.150 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.down_proj_weight_quantizers.151 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.down_proj_weight_quantizers.152 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.down_proj_weight_quantizers.153 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.down_proj_weight_quantizers.154 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.16e-01 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.down_proj_weight_quantizers.155 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.down_proj_weight_quantizers.156 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.down_proj_weight_quantizers.157 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.27e-01 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.down_proj_weight_quantizers.158 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.down_proj_weight_quantizers.159 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.down_proj_weight_quantizers.160 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.down_proj_weight_quantizers.161 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.14e-01 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.down_proj_weight_quantizers.162 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.down_proj_weight_quantizers.163 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.17e-01 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.down_proj_weight_quantizers.164 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.41e-01 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.down_proj_weight_quantizers.165 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.99e-01 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.down_proj_weight_quantizers.166 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.down_proj_weight_quantizers.167 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.down_proj_weight_quantizers.168 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.down_proj_weight_quantizers.169 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.down_proj_weight_quantizers.170 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.down_proj_weight_quantizers.171 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.00e-01 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.down_proj_weight_quantizers.172 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.59e-02 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.down_proj_weight_quantizers.173 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.21e-01 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.down_proj_weight_quantizers.174 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.down_proj_weight_quantizers.175 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.06e-01 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.down_proj_weight_quantizers.176 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.down_proj_weight_quantizers.177 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.34e-01 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.down_proj_weight_quantizers.178 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.46e-01 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.down_proj_weight_quantizers.179 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.39e-01 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.down_proj_weight_quantizers.180 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.down_proj_weight_quantizers.181 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.down_proj_weight_quantizers.182 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.down_proj_weight_quantizers.183 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.down_proj_weight_quantizers.184 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.27e-01 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.down_proj_weight_quantizers.185 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.down_proj_weight_quantizers.186 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.down_proj_weight_quantizers.187 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.down_proj_weight_quantizers.188 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.down_proj_weight_quantizers.189 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.20e-01 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.down_proj_weight_quantizers.190 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.down_proj_weight_quantizers.191 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.down_proj_weight_quantizers.192 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.down_proj_weight_quantizers.193 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.down_proj_weight_quantizers.194 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.67e-01 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.down_proj_weight_quantizers.195 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.down_proj_weight_quantizers.196 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.down_proj_weight_quantizers.197 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.down_proj_weight_quantizers.198 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.down_proj_weight_quantizers.199 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.25e-01 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.down_proj_weight_quantizers.200 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.down_proj_weight_quantizers.201 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.down_proj_weight_quantizers.202 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.down_proj_weight_quantizers.203 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.down_proj_weight_quantizers.204 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.down_proj_weight_quantizers.205 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.down_proj_weight_quantizers.206 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.down_proj_weight_quantizers.207 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.down_proj_weight_quantizers.208 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.down_proj_weight_quantizers.209 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.59e-02 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.down_proj_weight_quantizers.210 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.down_proj_weight_quantizers.211 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.down_proj_weight_quantizers.212 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.down_proj_weight_quantizers.213 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.down_proj_weight_quantizers.214 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.52e-01 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.down_proj_weight_quantizers.215 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.down_proj_weight_quantizers.216 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.50e-02 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.down_proj_weight_quantizers.217 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.15e-01 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.down_proj_weight_quantizers.218 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.down_proj_weight_quantizers.219 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.down_proj_weight_quantizers.220 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.down_proj_weight_quantizers.221 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.down_proj_weight_quantizers.222 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.26e-01 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.down_proj_weight_quantizers.223 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.down_proj_weight_quantizers.224 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.27e-01 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.down_proj_weight_quantizers.225 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.down_proj_weight_quantizers.226 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.15e-02 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.down_proj_weight_quantizers.227 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.down_proj_weight_quantizers.228 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.down_proj_weight_quantizers.229 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.down_proj_weight_quantizers.230 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.down_proj_weight_quantizers.231 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.down_proj_weight_quantizers.232 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.down_proj_weight_quantizers.233 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.19e-01 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.down_proj_weight_quantizers.234 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.50e-02 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.down_proj_weight_quantizers.235 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.75e-01 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.down_proj_weight_quantizers.236 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.64e-02 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.down_proj_weight_quantizers.237 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.down_proj_weight_quantizers.238 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.down_proj_weight_quantizers.239 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.down_proj_weight_quantizers.240 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.15e-01 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.down_proj_weight_quantizers.241 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.down_proj_weight_quantizers.242 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.down_proj_weight_quantizers.243 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.49e-01 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.down_proj_weight_quantizers.244 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.down_proj_weight_quantizers.245 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.down_proj_weight_quantizers.246 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.down_proj_weight_quantizers.247 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.50e-02 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.down_proj_weight_quantizers.248 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.down_proj_weight_quantizers.249 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.down_proj_weight_quantizers.250 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.54e-02 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.down_proj_weight_quantizers.251 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.down_proj_weight_quantizers.252 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.down_proj_weight_quantizers.253 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.down_proj_weight_quantizers.254 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.down_proj_weight_quantizers.255 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.50e-02 calibrator=MaxCalibrator quant) +model.layers.36.mlp.shared_experts.gate_proj.input_quantizer TensorQuantizer(disabled) +model.layers.36.mlp.shared_experts.gate_proj.output_quantizer TensorQuantizer(disabled) +model.layers.36.mlp.shared_experts.gate_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.36.mlp.shared_experts.up_proj.input_quantizer TensorQuantizer(disabled) +model.layers.36.mlp.shared_experts.up_proj.output_quantizer TensorQuantizer(disabled) +model.layers.36.mlp.shared_experts.up_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.36.mlp.shared_experts.down_proj.input_quantizer TensorQuantizer(disabled) +model.layers.36.mlp.shared_experts.down_proj.output_quantizer TensorQuantizer(disabled) +model.layers.36.mlp.shared_experts.down_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.37.self_attn.q_a_proj.input_quantizer TensorQuantizer(disabled) +model.layers.37.self_attn.q_a_proj.output_quantizer TensorQuantizer(disabled) +model.layers.37.self_attn.q_a_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.37.self_attn.q_b_proj.input_quantizer TensorQuantizer(disabled) +model.layers.37.self_attn.q_b_proj.output_quantizer TensorQuantizer(disabled) +model.layers.37.self_attn.q_b_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.37.self_attn.kv_a_proj_with_mqa.input_quantizer TensorQuantizer(disabled) +model.layers.37.self_attn.kv_a_proj_with_mqa.output_quantizer TensorQuantizer(disabled) +model.layers.37.self_attn.kv_a_proj_with_mqa.weight_quantizer TensorQuantizer(disabled) +model.layers.37.self_attn.kv_b_proj.input_quantizer TensorQuantizer(disabled) +model.layers.37.self_attn.kv_b_proj.output_quantizer TensorQuantizer(disabled) +model.layers.37.self_attn.kv_b_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.37.self_attn.o_proj.input_quantizer TensorQuantizer(disabled) +model.layers.37.self_attn.o_proj.output_quantizer TensorQuantizer(disabled) +model.layers.37.self_attn.o_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.37.self_attn.q_bmm_quantizer TensorQuantizer(disabled) +model.layers.37.self_attn.k_bmm_quantizer TensorQuantizer((4, 3) bit fake per-tensor amax=4.48e+02(const) calibrator=MaxCalibrator quant) +model.layers.37.self_attn.v_bmm_quantizer TensorQuantizer((4, 3) bit fake per-tensor amax=4.48e+02(const) calibrator=MaxCalibrator quant) +model.layers.37.self_attn.p_bmm_quantizer TensorQuantizer(disabled) +model.layers.37.mlp.experts.gate_up_proj_input_quantizer TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.70e+00 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.gate_up_proj_weight_quantizers.0 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.22e-01 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.gate_up_proj_weight_quantizers.1 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.35e-01 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.gate_up_proj_weight_quantizers.2 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.gate_up_proj_weight_quantizers.3 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.gate_up_proj_weight_quantizers.4 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=5.78e-01 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.gate_up_proj_weight_quantizers.5 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.gate_up_proj_weight_quantizers.6 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.gate_up_proj_weight_quantizers.7 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.gate_up_proj_weight_quantizers.8 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.gate_up_proj_weight_quantizers.9 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.20e-01 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.gate_up_proj_weight_quantizers.10 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.gate_up_proj_weight_quantizers.11 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.gate_up_proj_weight_quantizers.12 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.gate_up_proj_weight_quantizers.13 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.gate_up_proj_weight_quantizers.14 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.45e-01 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.gate_up_proj_weight_quantizers.15 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.47e-01 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.gate_up_proj_weight_quantizers.16 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.56e-01 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.gate_up_proj_weight_quantizers.17 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.36e-01 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.gate_up_proj_weight_quantizers.18 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.gate_up_proj_weight_quantizers.19 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.gate_up_proj_weight_quantizers.20 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.58e-01 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.gate_up_proj_weight_quantizers.21 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.37e-01 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.gate_up_proj_weight_quantizers.22 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.gate_up_proj_weight_quantizers.23 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.gate_up_proj_weight_quantizers.24 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.gate_up_proj_weight_quantizers.25 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.33e-01 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.gate_up_proj_weight_quantizers.26 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.gate_up_proj_weight_quantizers.27 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.gate_up_proj_weight_quantizers.28 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.gate_up_proj_weight_quantizers.29 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.37e-01 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.gate_up_proj_weight_quantizers.30 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.08e-01 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.gate_up_proj_weight_quantizers.31 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.gate_up_proj_weight_quantizers.32 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.16e-01 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.gate_up_proj_weight_quantizers.33 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.gate_up_proj_weight_quantizers.34 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.43e-01 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.gate_up_proj_weight_quantizers.35 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.64e-02 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.gate_up_proj_weight_quantizers.36 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.gate_up_proj_weight_quantizers.37 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.gate_up_proj_weight_quantizers.38 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.gate_up_proj_weight_quantizers.39 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.gate_up_proj_weight_quantizers.40 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.gate_up_proj_weight_quantizers.41 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.gate_up_proj_weight_quantizers.42 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.gate_up_proj_weight_quantizers.43 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.68e-01 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.gate_up_proj_weight_quantizers.44 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.40e-01 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.gate_up_proj_weight_quantizers.45 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.gate_up_proj_weight_quantizers.46 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.gate_up_proj_weight_quantizers.47 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.61e-01 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.gate_up_proj_weight_quantizers.48 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.gate_up_proj_weight_quantizers.49 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=4.80e-01 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.gate_up_proj_weight_quantizers.50 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.56e-01 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.gate_up_proj_weight_quantizers.51 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.gate_up_proj_weight_quantizers.52 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.gate_up_proj_weight_quantizers.53 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.gate_up_proj_weight_quantizers.54 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.gate_up_proj_weight_quantizers.55 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.gate_up_proj_weight_quantizers.56 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.25e-01 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.gate_up_proj_weight_quantizers.57 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.28e-01 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.gate_up_proj_weight_quantizers.58 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.46e-01 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.gate_up_proj_weight_quantizers.59 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.29e-01 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.gate_up_proj_weight_quantizers.60 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.79e-01 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.gate_up_proj_weight_quantizers.61 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.gate_up_proj_weight_quantizers.62 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.gate_up_proj_weight_quantizers.63 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.42e-01 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.gate_up_proj_weight_quantizers.64 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.gate_up_proj_weight_quantizers.65 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.gate_up_proj_weight_quantizers.66 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.gate_up_proj_weight_quantizers.67 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.gate_up_proj_weight_quantizers.68 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.gate_up_proj_weight_quantizers.69 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.gate_up_proj_weight_quantizers.70 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.gate_up_proj_weight_quantizers.71 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.28e-01 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.gate_up_proj_weight_quantizers.72 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.99e-01 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.gate_up_proj_weight_quantizers.73 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.20e-01 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.gate_up_proj_weight_quantizers.74 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.30e-01 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.gate_up_proj_weight_quantizers.75 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.gate_up_proj_weight_quantizers.76 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.gate_up_proj_weight_quantizers.77 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.73e-01 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.gate_up_proj_weight_quantizers.78 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.23e-01 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.gate_up_proj_weight_quantizers.79 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.58e-01 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.gate_up_proj_weight_quantizers.80 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.gate_up_proj_weight_quantizers.81 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.90e-01 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.gate_up_proj_weight_quantizers.82 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.78e-01 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.gate_up_proj_weight_quantizers.83 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.gate_up_proj_weight_quantizers.84 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.gate_up_proj_weight_quantizers.85 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.gate_up_proj_weight_quantizers.86 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.69e-01 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.gate_up_proj_weight_quantizers.87 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.25e-01 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.gate_up_proj_weight_quantizers.88 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.14e-01 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.gate_up_proj_weight_quantizers.89 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.46e-01 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.gate_up_proj_weight_quantizers.90 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.80e-01 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.gate_up_proj_weight_quantizers.91 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.gate_up_proj_weight_quantizers.92 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.gate_up_proj_weight_quantizers.93 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.44e-01 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.gate_up_proj_weight_quantizers.94 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.gate_up_proj_weight_quantizers.95 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.gate_up_proj_weight_quantizers.96 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.gate_up_proj_weight_quantizers.97 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.gate_up_proj_weight_quantizers.98 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.33e-01 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.gate_up_proj_weight_quantizers.99 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.gate_up_proj_weight_quantizers.100 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.gate_up_proj_weight_quantizers.101 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.gate_up_proj_weight_quantizers.102 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.gate_up_proj_weight_quantizers.103 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.gate_up_proj_weight_quantizers.104 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.gate_up_proj_weight_quantizers.105 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.gate_up_proj_weight_quantizers.106 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.59e-01 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.gate_up_proj_weight_quantizers.107 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.63e-01 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.gate_up_proj_weight_quantizers.108 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.gate_up_proj_weight_quantizers.109 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.gate_up_proj_weight_quantizers.110 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.23e-01 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.gate_up_proj_weight_quantizers.111 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.gate_up_proj_weight_quantizers.112 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.96e-01 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.gate_up_proj_weight_quantizers.113 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.89e-01 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.gate_up_proj_weight_quantizers.114 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.21e-01 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.gate_up_proj_weight_quantizers.115 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.77e-01 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.gate_up_proj_weight_quantizers.116 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.32e-01 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.gate_up_proj_weight_quantizers.117 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.gate_up_proj_weight_quantizers.118 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.gate_up_proj_weight_quantizers.119 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.27e-01 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.gate_up_proj_weight_quantizers.120 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.16e-01 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.gate_up_proj_weight_quantizers.121 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.gate_up_proj_weight_quantizers.122 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.gate_up_proj_weight_quantizers.123 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.gate_up_proj_weight_quantizers.124 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.43e-01 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.gate_up_proj_weight_quantizers.125 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.22e-01 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.gate_up_proj_weight_quantizers.126 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.13e-01 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.gate_up_proj_weight_quantizers.127 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.gate_up_proj_weight_quantizers.128 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.23e-01 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.gate_up_proj_weight_quantizers.129 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.75e-01 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.gate_up_proj_weight_quantizers.130 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.58e-01 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.gate_up_proj_weight_quantizers.131 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.gate_up_proj_weight_quantizers.132 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.gate_up_proj_weight_quantizers.133 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.97e-01 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.gate_up_proj_weight_quantizers.134 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.13e-01 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.gate_up_proj_weight_quantizers.135 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.gate_up_proj_weight_quantizers.136 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.gate_up_proj_weight_quantizers.137 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.gate_up_proj_weight_quantizers.138 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.gate_up_proj_weight_quantizers.139 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.gate_up_proj_weight_quantizers.140 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.14e-01 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.gate_up_proj_weight_quantizers.141 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.35e-01 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.gate_up_proj_weight_quantizers.142 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.gate_up_proj_weight_quantizers.143 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.13e-01 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.gate_up_proj_weight_quantizers.144 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.gate_up_proj_weight_quantizers.145 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.gate_up_proj_weight_quantizers.146 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.46e-01 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.gate_up_proj_weight_quantizers.147 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.gate_up_proj_weight_quantizers.148 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.gate_up_proj_weight_quantizers.149 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.gate_up_proj_weight_quantizers.150 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.gate_up_proj_weight_quantizers.151 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.gate_up_proj_weight_quantizers.152 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.gate_up_proj_weight_quantizers.153 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.60e-01 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.gate_up_proj_weight_quantizers.154 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.79e-01 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.gate_up_proj_weight_quantizers.155 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.26e-01 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.gate_up_proj_weight_quantizers.156 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.gate_up_proj_weight_quantizers.157 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.gate_up_proj_weight_quantizers.158 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.gate_up_proj_weight_quantizers.159 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.gate_up_proj_weight_quantizers.160 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.gate_up_proj_weight_quantizers.161 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.gate_up_proj_weight_quantizers.162 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.24e-01 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.gate_up_proj_weight_quantizers.163 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.gate_up_proj_weight_quantizers.164 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.03e-01 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.gate_up_proj_weight_quantizers.165 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.gate_up_proj_weight_quantizers.166 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.gate_up_proj_weight_quantizers.167 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.44e-01 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.gate_up_proj_weight_quantizers.168 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.21e-01 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.gate_up_proj_weight_quantizers.169 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.gate_up_proj_weight_quantizers.170 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.gate_up_proj_weight_quantizers.171 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.gate_up_proj_weight_quantizers.172 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.gate_up_proj_weight_quantizers.173 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.gate_up_proj_weight_quantizers.174 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.gate_up_proj_weight_quantizers.175 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.gate_up_proj_weight_quantizers.176 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.gate_up_proj_weight_quantizers.177 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.gate_up_proj_weight_quantizers.178 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.57e-01 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.gate_up_proj_weight_quantizers.179 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.gate_up_proj_weight_quantizers.180 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.20e-01 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.gate_up_proj_weight_quantizers.181 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.60e-01 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.gate_up_proj_weight_quantizers.182 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.gate_up_proj_weight_quantizers.183 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.gate_up_proj_weight_quantizers.184 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.96e-01 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.gate_up_proj_weight_quantizers.185 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.69e-01 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.gate_up_proj_weight_quantizers.186 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.46e-01 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.gate_up_proj_weight_quantizers.187 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.gate_up_proj_weight_quantizers.188 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.gate_up_proj_weight_quantizers.189 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.38e-01 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.gate_up_proj_weight_quantizers.190 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.24e-01 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.gate_up_proj_weight_quantizers.191 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.gate_up_proj_weight_quantizers.192 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.50e-01 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.gate_up_proj_weight_quantizers.193 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.gate_up_proj_weight_quantizers.194 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.gate_up_proj_weight_quantizers.195 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.16e-01 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.gate_up_proj_weight_quantizers.196 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.gate_up_proj_weight_quantizers.197 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.gate_up_proj_weight_quantizers.198 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.26e-01 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.gate_up_proj_weight_quantizers.199 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.gate_up_proj_weight_quantizers.200 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.gate_up_proj_weight_quantizers.201 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.gate_up_proj_weight_quantizers.202 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.96e-01 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.gate_up_proj_weight_quantizers.203 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.gate_up_proj_weight_quantizers.204 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.gate_up_proj_weight_quantizers.205 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.gate_up_proj_weight_quantizers.206 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.gate_up_proj_weight_quantizers.207 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.gate_up_proj_weight_quantizers.208 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.20e-01 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.gate_up_proj_weight_quantizers.209 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.gate_up_proj_weight_quantizers.210 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.47e-01 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.gate_up_proj_weight_quantizers.211 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.64e-02 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.gate_up_proj_weight_quantizers.212 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.gate_up_proj_weight_quantizers.213 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.gate_up_proj_weight_quantizers.214 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.gate_up_proj_weight_quantizers.215 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.gate_up_proj_weight_quantizers.216 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.gate_up_proj_weight_quantizers.217 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.gate_up_proj_weight_quantizers.218 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.gate_up_proj_weight_quantizers.219 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.19e-01 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.gate_up_proj_weight_quantizers.220 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.40e-01 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.gate_up_proj_weight_quantizers.221 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.38e-01 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.gate_up_proj_weight_quantizers.222 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.30e-01 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.gate_up_proj_weight_quantizers.223 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.22e-01 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.gate_up_proj_weight_quantizers.224 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.47e-01 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.gate_up_proj_weight_quantizers.225 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.gate_up_proj_weight_quantizers.226 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.gate_up_proj_weight_quantizers.227 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.93e-01 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.gate_up_proj_weight_quantizers.228 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.gate_up_proj_weight_quantizers.229 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.gate_up_proj_weight_quantizers.230 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.52e-01 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.gate_up_proj_weight_quantizers.231 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.24e-01 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.gate_up_proj_weight_quantizers.232 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.gate_up_proj_weight_quantizers.233 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=4.86e-01 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.gate_up_proj_weight_quantizers.234 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.gate_up_proj_weight_quantizers.235 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.gate_up_proj_weight_quantizers.236 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.gate_up_proj_weight_quantizers.237 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.gate_up_proj_weight_quantizers.238 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.gate_up_proj_weight_quantizers.239 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.gate_up_proj_weight_quantizers.240 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.gate_up_proj_weight_quantizers.241 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.65e-01 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.gate_up_proj_weight_quantizers.242 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.gate_up_proj_weight_quantizers.243 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.01e-01 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.gate_up_proj_weight_quantizers.244 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.gate_up_proj_weight_quantizers.245 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.gate_up_proj_weight_quantizers.246 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.38e-01 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.gate_up_proj_weight_quantizers.247 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.gate_up_proj_weight_quantizers.248 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.13e-01 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.gate_up_proj_weight_quantizers.249 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.gate_up_proj_weight_quantizers.250 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.gate_up_proj_weight_quantizers.251 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.gate_up_proj_weight_quantizers.252 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.gate_up_proj_weight_quantizers.253 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.gate_up_proj_weight_quantizers.254 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.gate_up_proj_weight_quantizers.255 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.down_proj_input_quantizer TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.41e+01 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.down_proj_weight_quantizers.0 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.down_proj_weight_quantizers.1 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.down_proj_weight_quantizers.2 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.45e-02 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.down_proj_weight_quantizers.3 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.down_proj_weight_quantizers.4 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.61e-01 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.down_proj_weight_quantizers.5 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.down_proj_weight_quantizers.6 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.down_proj_weight_quantizers.7 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.down_proj_weight_quantizers.8 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.down_proj_weight_quantizers.9 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.down_proj_weight_quantizers.10 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.down_proj_weight_quantizers.11 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.down_proj_weight_quantizers.12 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.down_proj_weight_quantizers.13 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.down_proj_weight_quantizers.14 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.down_proj_weight_quantizers.15 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.37e-01 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.down_proj_weight_quantizers.16 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.down_proj_weight_quantizers.17 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.down_proj_weight_quantizers.18 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.down_proj_weight_quantizers.19 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.down_proj_weight_quantizers.20 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.34e-01 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.down_proj_weight_quantizers.21 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.down_proj_weight_quantizers.22 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.down_proj_weight_quantizers.23 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.down_proj_weight_quantizers.24 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.down_proj_weight_quantizers.25 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.down_proj_weight_quantizers.26 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.down_proj_weight_quantizers.27 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.down_proj_weight_quantizers.28 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.down_proj_weight_quantizers.29 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.down_proj_weight_quantizers.30 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.67e-01 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.down_proj_weight_quantizers.31 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.down_proj_weight_quantizers.32 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.down_proj_weight_quantizers.33 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.down_proj_weight_quantizers.34 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.down_proj_weight_quantizers.35 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.down_proj_weight_quantizers.36 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.down_proj_weight_quantizers.37 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.down_proj_weight_quantizers.38 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.down_proj_weight_quantizers.39 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.35e-02 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.down_proj_weight_quantizers.40 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.down_proj_weight_quantizers.41 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.50e-02 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.down_proj_weight_quantizers.42 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.down_proj_weight_quantizers.43 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.down_proj_weight_quantizers.44 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.down_proj_weight_quantizers.45 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.down_proj_weight_quantizers.46 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.down_proj_weight_quantizers.47 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.down_proj_weight_quantizers.48 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.down_proj_weight_quantizers.49 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.62e-01 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.down_proj_weight_quantizers.50 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.21e-01 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.down_proj_weight_quantizers.51 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.down_proj_weight_quantizers.52 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.40e-02 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.down_proj_weight_quantizers.53 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.45e-02 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.down_proj_weight_quantizers.54 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.down_proj_weight_quantizers.55 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.down_proj_weight_quantizers.56 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.54e-02 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.down_proj_weight_quantizers.57 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.down_proj_weight_quantizers.58 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.64e-02 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.down_proj_weight_quantizers.59 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.down_proj_weight_quantizers.60 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.88e-01 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.down_proj_weight_quantizers.61 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.down_proj_weight_quantizers.62 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.down_proj_weight_quantizers.63 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.25e-01 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.down_proj_weight_quantizers.64 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.66e-01 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.down_proj_weight_quantizers.65 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.down_proj_weight_quantizers.66 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.down_proj_weight_quantizers.67 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.33e-01 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.down_proj_weight_quantizers.68 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.down_proj_weight_quantizers.69 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.down_proj_weight_quantizers.70 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.down_proj_weight_quantizers.71 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.81e-01 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.down_proj_weight_quantizers.72 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=4.71e-01 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.down_proj_weight_quantizers.73 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.26e-01 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.down_proj_weight_quantizers.74 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.down_proj_weight_quantizers.75 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.down_proj_weight_quantizers.76 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.down_proj_weight_quantizers.77 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.down_proj_weight_quantizers.78 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.down_proj_weight_quantizers.79 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.down_proj_weight_quantizers.80 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.down_proj_weight_quantizers.81 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.down_proj_weight_quantizers.82 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.down_proj_weight_quantizers.83 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.down_proj_weight_quantizers.84 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.down_proj_weight_quantizers.85 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.down_proj_weight_quantizers.86 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.down_proj_weight_quantizers.87 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.down_proj_weight_quantizers.88 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.down_proj_weight_quantizers.89 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.54e-02 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.down_proj_weight_quantizers.90 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.35e-01 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.down_proj_weight_quantizers.91 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.down_proj_weight_quantizers.92 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.down_proj_weight_quantizers.93 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.down_proj_weight_quantizers.94 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.down_proj_weight_quantizers.95 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.down_proj_weight_quantizers.96 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.down_proj_weight_quantizers.97 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.64e-02 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.down_proj_weight_quantizers.98 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.down_proj_weight_quantizers.99 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.40e-02 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.down_proj_weight_quantizers.100 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.down_proj_weight_quantizers.101 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.down_proj_weight_quantizers.102 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.down_proj_weight_quantizers.103 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.64e-02 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.down_proj_weight_quantizers.104 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.down_proj_weight_quantizers.105 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.50e-02 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.down_proj_weight_quantizers.106 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.down_proj_weight_quantizers.107 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.down_proj_weight_quantizers.108 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.down_proj_weight_quantizers.109 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.54e-02 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.down_proj_weight_quantizers.110 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.down_proj_weight_quantizers.111 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.59e-02 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.down_proj_weight_quantizers.112 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.down_proj_weight_quantizers.113 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.16e-01 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.down_proj_weight_quantizers.114 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.down_proj_weight_quantizers.115 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.18e-01 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.down_proj_weight_quantizers.116 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.down_proj_weight_quantizers.117 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.down_proj_weight_quantizers.118 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.down_proj_weight_quantizers.119 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.down_proj_weight_quantizers.120 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.64e-02 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.down_proj_weight_quantizers.121 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.down_proj_weight_quantizers.122 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.50e-02 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.down_proj_weight_quantizers.123 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.45e-02 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.down_proj_weight_quantizers.124 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.down_proj_weight_quantizers.125 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.17e-01 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.down_proj_weight_quantizers.126 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.down_proj_weight_quantizers.127 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.down_proj_weight_quantizers.128 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.down_proj_weight_quantizers.129 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.down_proj_weight_quantizers.130 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.down_proj_weight_quantizers.131 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.down_proj_weight_quantizers.132 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.down_proj_weight_quantizers.133 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.73e-01 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.down_proj_weight_quantizers.134 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.52e-01 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.down_proj_weight_quantizers.135 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.down_proj_weight_quantizers.136 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.down_proj_weight_quantizers.137 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.59e-02 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.down_proj_weight_quantizers.138 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.down_proj_weight_quantizers.139 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.down_proj_weight_quantizers.140 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.down_proj_weight_quantizers.141 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.down_proj_weight_quantizers.142 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.down_proj_weight_quantizers.143 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.34e-01 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.down_proj_weight_quantizers.144 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.down_proj_weight_quantizers.145 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.down_proj_weight_quantizers.146 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.down_proj_weight_quantizers.147 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.down_proj_weight_quantizers.148 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.down_proj_weight_quantizers.149 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.down_proj_weight_quantizers.150 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.64e-02 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.down_proj_weight_quantizers.151 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.down_proj_weight_quantizers.152 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.down_proj_weight_quantizers.153 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.down_proj_weight_quantizers.154 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.16e-01 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.down_proj_weight_quantizers.155 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.64e-02 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.down_proj_weight_quantizers.156 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.down_proj_weight_quantizers.157 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.down_proj_weight_quantizers.158 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.down_proj_weight_quantizers.159 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.down_proj_weight_quantizers.160 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.64e-02 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.down_proj_weight_quantizers.161 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.down_proj_weight_quantizers.162 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.down_proj_weight_quantizers.163 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.down_proj_weight_quantizers.164 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.43e-01 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.down_proj_weight_quantizers.165 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.down_proj_weight_quantizers.166 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.down_proj_weight_quantizers.167 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.34e-01 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.down_proj_weight_quantizers.168 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.25e-02 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.down_proj_weight_quantizers.169 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.down_proj_weight_quantizers.170 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.down_proj_weight_quantizers.171 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.59e-02 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.down_proj_weight_quantizers.172 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.down_proj_weight_quantizers.173 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.down_proj_weight_quantizers.174 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.down_proj_weight_quantizers.175 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.50e-02 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.down_proj_weight_quantizers.176 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.64e-02 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.down_proj_weight_quantizers.177 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.down_proj_weight_quantizers.178 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.down_proj_weight_quantizers.179 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.down_proj_weight_quantizers.180 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.down_proj_weight_quantizers.181 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.down_proj_weight_quantizers.182 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.down_proj_weight_quantizers.183 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.down_proj_weight_quantizers.184 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.down_proj_weight_quantizers.185 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.25e-01 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.down_proj_weight_quantizers.186 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.down_proj_weight_quantizers.187 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.down_proj_weight_quantizers.188 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.down_proj_weight_quantizers.189 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.down_proj_weight_quantizers.190 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.65e-01 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.down_proj_weight_quantizers.191 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.down_proj_weight_quantizers.192 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.down_proj_weight_quantizers.193 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.down_proj_weight_quantizers.194 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.down_proj_weight_quantizers.195 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.down_proj_weight_quantizers.196 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.down_proj_weight_quantizers.197 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.down_proj_weight_quantizers.198 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.down_proj_weight_quantizers.199 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.54e-02 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.down_proj_weight_quantizers.200 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.down_proj_weight_quantizers.201 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.down_proj_weight_quantizers.202 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.down_proj_weight_quantizers.203 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.down_proj_weight_quantizers.204 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.down_proj_weight_quantizers.205 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.down_proj_weight_quantizers.206 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.down_proj_weight_quantizers.207 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.down_proj_weight_quantizers.208 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.down_proj_weight_quantizers.209 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.down_proj_weight_quantizers.210 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.64e-02 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.down_proj_weight_quantizers.211 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.down_proj_weight_quantizers.212 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.54e-02 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.down_proj_weight_quantizers.213 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.22e-01 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.down_proj_weight_quantizers.214 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.down_proj_weight_quantizers.215 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.down_proj_weight_quantizers.216 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.down_proj_weight_quantizers.217 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.down_proj_weight_quantizers.218 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.down_proj_weight_quantizers.219 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.28e-01 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.down_proj_weight_quantizers.220 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.down_proj_weight_quantizers.221 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.down_proj_weight_quantizers.222 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.down_proj_weight_quantizers.223 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.down_proj_weight_quantizers.224 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.64e-02 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.down_proj_weight_quantizers.225 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.down_proj_weight_quantizers.226 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.down_proj_weight_quantizers.227 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.41e-01 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.down_proj_weight_quantizers.228 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.down_proj_weight_quantizers.229 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.40e-02 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.down_proj_weight_quantizers.230 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.39e-01 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.down_proj_weight_quantizers.231 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.down_proj_weight_quantizers.232 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.down_proj_weight_quantizers.233 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.18e-01 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.down_proj_weight_quantizers.234 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.down_proj_weight_quantizers.235 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.down_proj_weight_quantizers.236 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.down_proj_weight_quantizers.237 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.down_proj_weight_quantizers.238 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.down_proj_weight_quantizers.239 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.down_proj_weight_quantizers.240 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.53e-01 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.down_proj_weight_quantizers.241 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.down_proj_weight_quantizers.242 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.down_proj_weight_quantizers.243 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.down_proj_weight_quantizers.244 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.down_proj_weight_quantizers.245 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.50e-02 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.down_proj_weight_quantizers.246 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.23e-01 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.down_proj_weight_quantizers.247 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.down_proj_weight_quantizers.248 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.down_proj_weight_quantizers.249 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.down_proj_weight_quantizers.250 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.down_proj_weight_quantizers.251 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.down_proj_weight_quantizers.252 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.64e-02 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.down_proj_weight_quantizers.253 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.down_proj_weight_quantizers.254 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.down_proj_weight_quantizers.255 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.37.mlp.shared_experts.gate_proj.input_quantizer TensorQuantizer(disabled) +model.layers.37.mlp.shared_experts.gate_proj.output_quantizer TensorQuantizer(disabled) +model.layers.37.mlp.shared_experts.gate_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.37.mlp.shared_experts.up_proj.input_quantizer TensorQuantizer(disabled) +model.layers.37.mlp.shared_experts.up_proj.output_quantizer TensorQuantizer(disabled) +model.layers.37.mlp.shared_experts.up_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.37.mlp.shared_experts.down_proj.input_quantizer TensorQuantizer(disabled) +model.layers.37.mlp.shared_experts.down_proj.output_quantizer TensorQuantizer(disabled) +model.layers.37.mlp.shared_experts.down_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.38.self_attn.q_a_proj.input_quantizer TensorQuantizer(disabled) +model.layers.38.self_attn.q_a_proj.output_quantizer TensorQuantizer(disabled) +model.layers.38.self_attn.q_a_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.38.self_attn.q_b_proj.input_quantizer TensorQuantizer(disabled) +model.layers.38.self_attn.q_b_proj.output_quantizer TensorQuantizer(disabled) +model.layers.38.self_attn.q_b_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.38.self_attn.kv_a_proj_with_mqa.input_quantizer TensorQuantizer(disabled) +model.layers.38.self_attn.kv_a_proj_with_mqa.output_quantizer TensorQuantizer(disabled) +model.layers.38.self_attn.kv_a_proj_with_mqa.weight_quantizer TensorQuantizer(disabled) +model.layers.38.self_attn.kv_b_proj.input_quantizer TensorQuantizer(disabled) +model.layers.38.self_attn.kv_b_proj.output_quantizer TensorQuantizer(disabled) +model.layers.38.self_attn.kv_b_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.38.self_attn.o_proj.input_quantizer TensorQuantizer(disabled) +model.layers.38.self_attn.o_proj.output_quantizer TensorQuantizer(disabled) +model.layers.38.self_attn.o_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.38.self_attn.indexer.wq_b.input_quantizer TensorQuantizer(disabled) +model.layers.38.self_attn.indexer.wq_b.output_quantizer TensorQuantizer(disabled) +model.layers.38.self_attn.indexer.wq_b.weight_quantizer TensorQuantizer(disabled) +model.layers.38.self_attn.indexer.wk.input_quantizer TensorQuantizer(disabled) +model.layers.38.self_attn.indexer.wk.output_quantizer TensorQuantizer(disabled) +model.layers.38.self_attn.indexer.wk.weight_quantizer TensorQuantizer(disabled) +model.layers.38.self_attn.indexer.k_norm.input_quantizer TensorQuantizer(disabled) +model.layers.38.self_attn.indexer.k_norm.output_quantizer TensorQuantizer(disabled) +model.layers.38.self_attn.indexer.weights_proj.input_quantizer TensorQuantizer(disabled) +model.layers.38.self_attn.indexer.weights_proj.output_quantizer TensorQuantizer(disabled) +model.layers.38.self_attn.indexer.weights_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.38.self_attn.q_bmm_quantizer TensorQuantizer(disabled) +model.layers.38.self_attn.k_bmm_quantizer TensorQuantizer((4, 3) bit fake per-tensor amax=4.48e+02(const) calibrator=MaxCalibrator quant) +model.layers.38.self_attn.v_bmm_quantizer TensorQuantizer((4, 3) bit fake per-tensor amax=4.48e+02(const) calibrator=MaxCalibrator quant) +model.layers.38.self_attn.p_bmm_quantizer TensorQuantizer(disabled) +model.layers.38.mlp.experts.gate_up_proj_input_quantizer TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.84e+00 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.gate_up_proj_weight_quantizers.0 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.gate_up_proj_weight_quantizers.1 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.gate_up_proj_weight_quantizers.2 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.gate_up_proj_weight_quantizers.3 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.19e-01 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.gate_up_proj_weight_quantizers.4 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.gate_up_proj_weight_quantizers.5 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.gate_up_proj_weight_quantizers.6 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.04e-01 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.gate_up_proj_weight_quantizers.7 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.13e-01 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.gate_up_proj_weight_quantizers.8 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.gate_up_proj_weight_quantizers.9 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.32e-01 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.gate_up_proj_weight_quantizers.10 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.01e-01 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.gate_up_proj_weight_quantizers.11 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.gate_up_proj_weight_quantizers.12 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.gate_up_proj_weight_quantizers.13 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.30e-01 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.gate_up_proj_weight_quantizers.14 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.gate_up_proj_weight_quantizers.15 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.37e-01 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.gate_up_proj_weight_quantizers.16 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.gate_up_proj_weight_quantizers.17 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.gate_up_proj_weight_quantizers.18 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.gate_up_proj_weight_quantizers.19 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.66e-01 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.gate_up_proj_weight_quantizers.20 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.48e-01 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.gate_up_proj_weight_quantizers.21 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.84e-01 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.gate_up_proj_weight_quantizers.22 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.gate_up_proj_weight_quantizers.23 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.22e-01 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.gate_up_proj_weight_quantizers.24 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.18e-01 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.gate_up_proj_weight_quantizers.25 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.43e-01 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.gate_up_proj_weight_quantizers.26 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.gate_up_proj_weight_quantizers.27 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.20e-01 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.gate_up_proj_weight_quantizers.28 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=7.07e-01 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.gate_up_proj_weight_quantizers.29 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.62e-01 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.gate_up_proj_weight_quantizers.30 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.44e-01 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.gate_up_proj_weight_quantizers.31 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.42e-01 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.gate_up_proj_weight_quantizers.32 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.gate_up_proj_weight_quantizers.33 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.gate_up_proj_weight_quantizers.34 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.45e-01 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.gate_up_proj_weight_quantizers.35 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.16e-01 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.gate_up_proj_weight_quantizers.36 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.gate_up_proj_weight_quantizers.37 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.gate_up_proj_weight_quantizers.38 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.gate_up_proj_weight_quantizers.39 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.72e-01 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.gate_up_proj_weight_quantizers.40 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.17e-01 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.gate_up_proj_weight_quantizers.41 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.24e-01 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.gate_up_proj_weight_quantizers.42 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.gate_up_proj_weight_quantizers.43 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.27e-01 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.gate_up_proj_weight_quantizers.44 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.gate_up_proj_weight_quantizers.45 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.81e-01 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.gate_up_proj_weight_quantizers.46 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.39e-01 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.gate_up_proj_weight_quantizers.47 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.20e-01 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.gate_up_proj_weight_quantizers.48 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.gate_up_proj_weight_quantizers.49 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.89e-01 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.gate_up_proj_weight_quantizers.50 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.gate_up_proj_weight_quantizers.51 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.57e-01 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.gate_up_proj_weight_quantizers.52 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.29e-01 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.gate_up_proj_weight_quantizers.53 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.47e-01 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.gate_up_proj_weight_quantizers.54 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.20e-01 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.gate_up_proj_weight_quantizers.55 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.15e-01 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.gate_up_proj_weight_quantizers.56 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.gate_up_proj_weight_quantizers.57 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.gate_up_proj_weight_quantizers.58 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.04e-01 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.gate_up_proj_weight_quantizers.59 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.42e-01 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.gate_up_proj_weight_quantizers.60 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.37e-01 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.gate_up_proj_weight_quantizers.61 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.gate_up_proj_weight_quantizers.62 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.18e-01 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.gate_up_proj_weight_quantizers.63 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.41e-01 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.gate_up_proj_weight_quantizers.64 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.48e-01 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.gate_up_proj_weight_quantizers.65 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.gate_up_proj_weight_quantizers.66 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.gate_up_proj_weight_quantizers.67 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.gate_up_proj_weight_quantizers.68 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.15e-01 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.gate_up_proj_weight_quantizers.69 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.gate_up_proj_weight_quantizers.70 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.gate_up_proj_weight_quantizers.71 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.gate_up_proj_weight_quantizers.72 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.gate_up_proj_weight_quantizers.73 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.94e-01 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.gate_up_proj_weight_quantizers.74 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.gate_up_proj_weight_quantizers.75 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.gate_up_proj_weight_quantizers.76 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.gate_up_proj_weight_quantizers.77 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.17e-01 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.gate_up_proj_weight_quantizers.78 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.gate_up_proj_weight_quantizers.79 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.gate_up_proj_weight_quantizers.80 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.gate_up_proj_weight_quantizers.81 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.gate_up_proj_weight_quantizers.82 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.13e-01 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.gate_up_proj_weight_quantizers.83 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.gate_up_proj_weight_quantizers.84 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.51e-01 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.gate_up_proj_weight_quantizers.85 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.28e-01 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.gate_up_proj_weight_quantizers.86 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.27e-01 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.gate_up_proj_weight_quantizers.87 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.31e-01 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.gate_up_proj_weight_quantizers.88 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.gate_up_proj_weight_quantizers.89 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.gate_up_proj_weight_quantizers.90 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.gate_up_proj_weight_quantizers.91 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.gate_up_proj_weight_quantizers.92 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.66e-01 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.gate_up_proj_weight_quantizers.93 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.42e-01 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.gate_up_proj_weight_quantizers.94 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.99e-01 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.gate_up_proj_weight_quantizers.95 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.gate_up_proj_weight_quantizers.96 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.gate_up_proj_weight_quantizers.97 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.gate_up_proj_weight_quantizers.98 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.53e-01 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.gate_up_proj_weight_quantizers.99 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.47e-01 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.gate_up_proj_weight_quantizers.100 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.gate_up_proj_weight_quantizers.101 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.47e-01 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.gate_up_proj_weight_quantizers.102 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.42e-01 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.gate_up_proj_weight_quantizers.103 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.gate_up_proj_weight_quantizers.104 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.43e-01 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.gate_up_proj_weight_quantizers.105 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.gate_up_proj_weight_quantizers.106 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.gate_up_proj_weight_quantizers.107 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.gate_up_proj_weight_quantizers.108 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.gate_up_proj_weight_quantizers.109 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.gate_up_proj_weight_quantizers.110 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.80e-01 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.gate_up_proj_weight_quantizers.111 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.gate_up_proj_weight_quantizers.112 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.gate_up_proj_weight_quantizers.113 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.44e-01 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.gate_up_proj_weight_quantizers.114 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.56e-01 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.gate_up_proj_weight_quantizers.115 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.gate_up_proj_weight_quantizers.116 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.gate_up_proj_weight_quantizers.117 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.gate_up_proj_weight_quantizers.118 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.gate_up_proj_weight_quantizers.119 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.gate_up_proj_weight_quantizers.120 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.53e-01 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.gate_up_proj_weight_quantizers.121 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.15e-01 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.gate_up_proj_weight_quantizers.122 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.33e-01 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.gate_up_proj_weight_quantizers.123 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.gate_up_proj_weight_quantizers.124 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.gate_up_proj_weight_quantizers.125 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.gate_up_proj_weight_quantizers.126 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.69e-01 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.gate_up_proj_weight_quantizers.127 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.gate_up_proj_weight_quantizers.128 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.17e-01 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.gate_up_proj_weight_quantizers.129 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.gate_up_proj_weight_quantizers.130 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.64e-01 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.gate_up_proj_weight_quantizers.131 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.gate_up_proj_weight_quantizers.132 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.gate_up_proj_weight_quantizers.133 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.25e-01 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.gate_up_proj_weight_quantizers.134 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.gate_up_proj_weight_quantizers.135 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.gate_up_proj_weight_quantizers.136 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.40e-01 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.gate_up_proj_weight_quantizers.137 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.gate_up_proj_weight_quantizers.138 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.gate_up_proj_weight_quantizers.139 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.gate_up_proj_weight_quantizers.140 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.gate_up_proj_weight_quantizers.141 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.00e-01 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.gate_up_proj_weight_quantizers.142 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.gate_up_proj_weight_quantizers.143 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.71e-01 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.gate_up_proj_weight_quantizers.144 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.24e-01 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.gate_up_proj_weight_quantizers.145 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.22e-01 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.gate_up_proj_weight_quantizers.146 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.13e-01 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.gate_up_proj_weight_quantizers.147 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.gate_up_proj_weight_quantizers.148 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.52e-01 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.gate_up_proj_weight_quantizers.149 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.gate_up_proj_weight_quantizers.150 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.gate_up_proj_weight_quantizers.151 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.05e-01 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.gate_up_proj_weight_quantizers.152 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.gate_up_proj_weight_quantizers.153 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.gate_up_proj_weight_quantizers.154 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.gate_up_proj_weight_quantizers.155 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.gate_up_proj_weight_quantizers.156 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.gate_up_proj_weight_quantizers.157 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.98e-01 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.gate_up_proj_weight_quantizers.158 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.gate_up_proj_weight_quantizers.159 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=5.16e-01 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.gate_up_proj_weight_quantizers.160 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.30e-01 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.gate_up_proj_weight_quantizers.161 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.15e-01 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.gate_up_proj_weight_quantizers.162 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.12e-01 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.gate_up_proj_weight_quantizers.163 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.gate_up_proj_weight_quantizers.164 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.gate_up_proj_weight_quantizers.165 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.gate_up_proj_weight_quantizers.166 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.gate_up_proj_weight_quantizers.167 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.17e-01 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.gate_up_proj_weight_quantizers.168 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.gate_up_proj_weight_quantizers.169 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.gate_up_proj_weight_quantizers.170 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.14e-01 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.gate_up_proj_weight_quantizers.171 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.gate_up_proj_weight_quantizers.172 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.03e-01 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.gate_up_proj_weight_quantizers.173 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.25e-01 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.gate_up_proj_weight_quantizers.174 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.gate_up_proj_weight_quantizers.175 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.19e-01 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.gate_up_proj_weight_quantizers.176 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.gate_up_proj_weight_quantizers.177 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.gate_up_proj_weight_quantizers.178 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.gate_up_proj_weight_quantizers.179 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.gate_up_proj_weight_quantizers.180 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.17e-01 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.gate_up_proj_weight_quantizers.181 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.gate_up_proj_weight_quantizers.182 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.30e-01 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.gate_up_proj_weight_quantizers.183 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.84e-01 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.gate_up_proj_weight_quantizers.184 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.gate_up_proj_weight_quantizers.185 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.97e-01 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.gate_up_proj_weight_quantizers.186 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.gate_up_proj_weight_quantizers.187 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.gate_up_proj_weight_quantizers.188 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.gate_up_proj_weight_quantizers.189 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.36e-01 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.gate_up_proj_weight_quantizers.190 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.55e-01 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.gate_up_proj_weight_quantizers.191 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.11e-01 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.gate_up_proj_weight_quantizers.192 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.31e-01 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.gate_up_proj_weight_quantizers.193 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.gate_up_proj_weight_quantizers.194 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.gate_up_proj_weight_quantizers.195 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.42e-01 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.gate_up_proj_weight_quantizers.196 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.gate_up_proj_weight_quantizers.197 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.gate_up_proj_weight_quantizers.198 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.gate_up_proj_weight_quantizers.199 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.23e-01 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.gate_up_proj_weight_quantizers.200 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.01e-01 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.gate_up_proj_weight_quantizers.201 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.62e-01 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.gate_up_proj_weight_quantizers.202 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.83e-01 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.gate_up_proj_weight_quantizers.203 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.46e-01 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.gate_up_proj_weight_quantizers.204 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.gate_up_proj_weight_quantizers.205 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.gate_up_proj_weight_quantizers.206 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.64e-01 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.gate_up_proj_weight_quantizers.207 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=5.94e-01 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.gate_up_proj_weight_quantizers.208 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.95e-01 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.gate_up_proj_weight_quantizers.209 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.gate_up_proj_weight_quantizers.210 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.gate_up_proj_weight_quantizers.211 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.gate_up_proj_weight_quantizers.212 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.gate_up_proj_weight_quantizers.213 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.00e-01 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.gate_up_proj_weight_quantizers.214 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.42e-01 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.gate_up_proj_weight_quantizers.215 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.41e-01 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.gate_up_proj_weight_quantizers.216 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=4.57e-01 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.gate_up_proj_weight_quantizers.217 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.gate_up_proj_weight_quantizers.218 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.88e-01 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.gate_up_proj_weight_quantizers.219 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.gate_up_proj_weight_quantizers.220 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.16e-01 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.gate_up_proj_weight_quantizers.221 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.46e-01 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.gate_up_proj_weight_quantizers.222 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.gate_up_proj_weight_quantizers.223 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.01e-01 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.gate_up_proj_weight_quantizers.224 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.98e-01 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.gate_up_proj_weight_quantizers.225 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.71e-01 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.gate_up_proj_weight_quantizers.226 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.16e-01 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.gate_up_proj_weight_quantizers.227 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.32e-01 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.gate_up_proj_weight_quantizers.228 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.gate_up_proj_weight_quantizers.229 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.gate_up_proj_weight_quantizers.230 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.gate_up_proj_weight_quantizers.231 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.gate_up_proj_weight_quantizers.232 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.gate_up_proj_weight_quantizers.233 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.24e-01 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.gate_up_proj_weight_quantizers.234 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.17e-01 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.gate_up_proj_weight_quantizers.235 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.gate_up_proj_weight_quantizers.236 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.70e-01 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.gate_up_proj_weight_quantizers.237 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.29e-01 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.gate_up_proj_weight_quantizers.238 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.gate_up_proj_weight_quantizers.239 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.70e-01 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.gate_up_proj_weight_quantizers.240 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.43e-01 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.gate_up_proj_weight_quantizers.241 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.20e-01 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.gate_up_proj_weight_quantizers.242 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.gate_up_proj_weight_quantizers.243 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.14e-01 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.gate_up_proj_weight_quantizers.244 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.gate_up_proj_weight_quantizers.245 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.gate_up_proj_weight_quantizers.246 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.19e-01 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.gate_up_proj_weight_quantizers.247 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.32e-01 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.gate_up_proj_weight_quantizers.248 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.gate_up_proj_weight_quantizers.249 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.83e-01 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.gate_up_proj_weight_quantizers.250 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.23e-01 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.gate_up_proj_weight_quantizers.251 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.gate_up_proj_weight_quantizers.252 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.27e-01 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.gate_up_proj_weight_quantizers.253 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.gate_up_proj_weight_quantizers.254 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.gate_up_proj_weight_quantizers.255 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.down_proj_input_quantizer TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.56e+01 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.down_proj_weight_quantizers.0 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.down_proj_weight_quantizers.1 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.down_proj_weight_quantizers.2 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.down_proj_weight_quantizers.3 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.down_proj_weight_quantizers.4 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.down_proj_weight_quantizers.5 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.down_proj_weight_quantizers.6 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.04e-01 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.down_proj_weight_quantizers.7 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.down_proj_weight_quantizers.8 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.down_proj_weight_quantizers.9 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.down_proj_weight_quantizers.10 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.95e-01 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.down_proj_weight_quantizers.11 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.down_proj_weight_quantizers.12 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.down_proj_weight_quantizers.13 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.down_proj_weight_quantizers.14 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.down_proj_weight_quantizers.15 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.down_proj_weight_quantizers.16 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.34e-01 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.down_proj_weight_quantizers.17 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.down_proj_weight_quantizers.18 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.40e-02 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.down_proj_weight_quantizers.19 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.15e-01 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.down_proj_weight_quantizers.20 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.down_proj_weight_quantizers.21 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.down_proj_weight_quantizers.22 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.down_proj_weight_quantizers.23 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.down_proj_weight_quantizers.24 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.59e-02 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.down_proj_weight_quantizers.25 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.97e-01 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.down_proj_weight_quantizers.26 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.down_proj_weight_quantizers.27 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.down_proj_weight_quantizers.28 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.53e-01 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.down_proj_weight_quantizers.29 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.down_proj_weight_quantizers.30 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.down_proj_weight_quantizers.31 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.down_proj_weight_quantizers.32 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.down_proj_weight_quantizers.33 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.down_proj_weight_quantizers.34 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.down_proj_weight_quantizers.35 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.down_proj_weight_quantizers.36 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.down_proj_weight_quantizers.37 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.down_proj_weight_quantizers.38 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.down_proj_weight_quantizers.39 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.31e-01 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.down_proj_weight_quantizers.40 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.down_proj_weight_quantizers.41 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.down_proj_weight_quantizers.42 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.50e-02 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.down_proj_weight_quantizers.43 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.down_proj_weight_quantizers.44 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.down_proj_weight_quantizers.45 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.83e-01 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.down_proj_weight_quantizers.46 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.down_proj_weight_quantizers.47 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.down_proj_weight_quantizers.48 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.down_proj_weight_quantizers.49 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.down_proj_weight_quantizers.50 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.64e-02 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.down_proj_weight_quantizers.51 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.down_proj_weight_quantizers.52 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.down_proj_weight_quantizers.53 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.down_proj_weight_quantizers.54 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.75e-01 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.down_proj_weight_quantizers.55 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.down_proj_weight_quantizers.56 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.down_proj_weight_quantizers.57 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.64e-02 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.down_proj_weight_quantizers.58 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.31e-01 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.down_proj_weight_quantizers.59 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.50e-02 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.down_proj_weight_quantizers.60 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.down_proj_weight_quantizers.61 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.down_proj_weight_quantizers.62 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.down_proj_weight_quantizers.63 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.54e-02 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.down_proj_weight_quantizers.64 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.down_proj_weight_quantizers.65 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.down_proj_weight_quantizers.66 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.down_proj_weight_quantizers.67 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.down_proj_weight_quantizers.68 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.down_proj_weight_quantizers.69 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.down_proj_weight_quantizers.70 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.down_proj_weight_quantizers.71 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.down_proj_weight_quantizers.72 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.down_proj_weight_quantizers.73 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.down_proj_weight_quantizers.74 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.59e-02 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.down_proj_weight_quantizers.75 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.50e-02 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.down_proj_weight_quantizers.76 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.down_proj_weight_quantizers.77 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.50e-02 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.down_proj_weight_quantizers.78 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.down_proj_weight_quantizers.79 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.down_proj_weight_quantizers.80 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.down_proj_weight_quantizers.81 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.down_proj_weight_quantizers.82 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.down_proj_weight_quantizers.83 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.50e-02 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.down_proj_weight_quantizers.84 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.down_proj_weight_quantizers.85 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.13e-01 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.down_proj_weight_quantizers.86 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.77e-01 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.down_proj_weight_quantizers.87 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.66e-01 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.down_proj_weight_quantizers.88 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.down_proj_weight_quantizers.89 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.down_proj_weight_quantizers.90 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.down_proj_weight_quantizers.91 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.down_proj_weight_quantizers.92 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.down_proj_weight_quantizers.93 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.52e-01 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.down_proj_weight_quantizers.94 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.down_proj_weight_quantizers.95 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.down_proj_weight_quantizers.96 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.49e-01 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.down_proj_weight_quantizers.97 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.down_proj_weight_quantizers.98 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.75e-01 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.down_proj_weight_quantizers.99 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.50e-01 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.down_proj_weight_quantizers.100 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.down_proj_weight_quantizers.101 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.down_proj_weight_quantizers.102 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.down_proj_weight_quantizers.103 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.down_proj_weight_quantizers.104 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.down_proj_weight_quantizers.105 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.59e-02 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.down_proj_weight_quantizers.106 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.down_proj_weight_quantizers.107 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.down_proj_weight_quantizers.108 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.down_proj_weight_quantizers.109 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.down_proj_weight_quantizers.110 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.down_proj_weight_quantizers.111 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.down_proj_weight_quantizers.112 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.down_proj_weight_quantizers.113 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.down_proj_weight_quantizers.114 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.30e-01 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.down_proj_weight_quantizers.115 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.64e-02 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.down_proj_weight_quantizers.116 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.down_proj_weight_quantizers.117 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.59e-02 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.down_proj_weight_quantizers.118 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.down_proj_weight_quantizers.119 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.down_proj_weight_quantizers.120 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.down_proj_weight_quantizers.121 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.down_proj_weight_quantizers.122 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.down_proj_weight_quantizers.123 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.down_proj_weight_quantizers.124 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.down_proj_weight_quantizers.125 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.down_proj_weight_quantizers.126 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.down_proj_weight_quantizers.127 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.down_proj_weight_quantizers.128 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.54e-01 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.down_proj_weight_quantizers.129 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.down_proj_weight_quantizers.130 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.89e-01 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.down_proj_weight_quantizers.131 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.40e-02 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.down_proj_weight_quantizers.132 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.down_proj_weight_quantizers.133 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.58e-01 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.down_proj_weight_quantizers.134 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.down_proj_weight_quantizers.135 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.down_proj_weight_quantizers.136 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.18e-01 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.down_proj_weight_quantizers.137 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.down_proj_weight_quantizers.138 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.down_proj_weight_quantizers.139 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.down_proj_weight_quantizers.140 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.down_proj_weight_quantizers.141 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.down_proj_weight_quantizers.142 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.down_proj_weight_quantizers.143 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.46e-01 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.down_proj_weight_quantizers.144 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.down_proj_weight_quantizers.145 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.40e-02 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.down_proj_weight_quantizers.146 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.down_proj_weight_quantizers.147 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.down_proj_weight_quantizers.148 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.27e-01 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.down_proj_weight_quantizers.149 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.down_proj_weight_quantizers.150 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.down_proj_weight_quantizers.151 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.down_proj_weight_quantizers.152 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.down_proj_weight_quantizers.153 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.down_proj_weight_quantizers.154 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.down_proj_weight_quantizers.155 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.down_proj_weight_quantizers.156 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.down_proj_weight_quantizers.157 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.down_proj_weight_quantizers.158 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.down_proj_weight_quantizers.159 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.27e-01 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.down_proj_weight_quantizers.160 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.down_proj_weight_quantizers.161 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.down_proj_weight_quantizers.162 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.down_proj_weight_quantizers.163 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.down_proj_weight_quantizers.164 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.50e-02 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.down_proj_weight_quantizers.165 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.down_proj_weight_quantizers.166 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.down_proj_weight_quantizers.167 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.17e-01 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.down_proj_weight_quantizers.168 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.down_proj_weight_quantizers.169 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.down_proj_weight_quantizers.170 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.20e-01 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.down_proj_weight_quantizers.171 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.down_proj_weight_quantizers.172 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.48e-01 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.down_proj_weight_quantizers.173 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.down_proj_weight_quantizers.174 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.64e-02 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.down_proj_weight_quantizers.175 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.27e-01 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.down_proj_weight_quantizers.176 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.down_proj_weight_quantizers.177 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.54e-02 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.down_proj_weight_quantizers.178 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.down_proj_weight_quantizers.179 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.down_proj_weight_quantizers.180 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.down_proj_weight_quantizers.181 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.down_proj_weight_quantizers.182 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.31e-01 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.down_proj_weight_quantizers.183 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.down_proj_weight_quantizers.184 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.down_proj_weight_quantizers.185 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.59e-01 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.down_proj_weight_quantizers.186 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.down_proj_weight_quantizers.187 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.down_proj_weight_quantizers.188 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.down_proj_weight_quantizers.189 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.down_proj_weight_quantizers.190 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.down_proj_weight_quantizers.191 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.59e-02 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.down_proj_weight_quantizers.192 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.38e-01 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.down_proj_weight_quantizers.193 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.down_proj_weight_quantizers.194 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.down_proj_weight_quantizers.195 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.down_proj_weight_quantizers.196 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.down_proj_weight_quantizers.197 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.down_proj_weight_quantizers.198 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.15e-01 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.down_proj_weight_quantizers.199 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.40e-02 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.down_proj_weight_quantizers.200 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.87e-01 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.down_proj_weight_quantizers.201 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.99e-01 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.down_proj_weight_quantizers.202 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.down_proj_weight_quantizers.203 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.30e-01 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.down_proj_weight_quantizers.204 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.down_proj_weight_quantizers.205 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.down_proj_weight_quantizers.206 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.down_proj_weight_quantizers.207 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.down_proj_weight_quantizers.208 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.44e-01 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.down_proj_weight_quantizers.209 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.down_proj_weight_quantizers.210 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.down_proj_weight_quantizers.211 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.down_proj_weight_quantizers.212 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.90e-01 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.down_proj_weight_quantizers.213 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.15e-01 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.down_proj_weight_quantizers.214 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.down_proj_weight_quantizers.215 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.down_proj_weight_quantizers.216 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.22e-01 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.down_proj_weight_quantizers.217 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.down_proj_weight_quantizers.218 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.64e-02 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.down_proj_weight_quantizers.219 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.down_proj_weight_quantizers.220 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.down_proj_weight_quantizers.221 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.down_proj_weight_quantizers.222 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.down_proj_weight_quantizers.223 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=4.73e-01 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.down_proj_weight_quantizers.224 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.down_proj_weight_quantizers.225 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.09e-01 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.down_proj_weight_quantizers.226 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.down_proj_weight_quantizers.227 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.down_proj_weight_quantizers.228 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.59e-02 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.down_proj_weight_quantizers.229 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.down_proj_weight_quantizers.230 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.down_proj_weight_quantizers.231 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.down_proj_weight_quantizers.232 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.down_proj_weight_quantizers.233 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.down_proj_weight_quantizers.234 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.54e-02 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.down_proj_weight_quantizers.235 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.down_proj_weight_quantizers.236 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.52e-01 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.down_proj_weight_quantizers.237 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.down_proj_weight_quantizers.238 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.down_proj_weight_quantizers.239 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.down_proj_weight_quantizers.240 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.down_proj_weight_quantizers.241 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.55e-01 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.down_proj_weight_quantizers.242 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.down_proj_weight_quantizers.243 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.down_proj_weight_quantizers.244 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.down_proj_weight_quantizers.245 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.down_proj_weight_quantizers.246 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.down_proj_weight_quantizers.247 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.35e-02 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.down_proj_weight_quantizers.248 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.down_proj_weight_quantizers.249 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.48e-01 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.down_proj_weight_quantizers.250 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.down_proj_weight_quantizers.251 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.down_proj_weight_quantizers.252 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.down_proj_weight_quantizers.253 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.down_proj_weight_quantizers.254 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.down_proj_weight_quantizers.255 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.38.mlp.shared_experts.gate_proj.input_quantizer TensorQuantizer(disabled) +model.layers.38.mlp.shared_experts.gate_proj.output_quantizer TensorQuantizer(disabled) +model.layers.38.mlp.shared_experts.gate_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.38.mlp.shared_experts.up_proj.input_quantizer TensorQuantizer(disabled) +model.layers.38.mlp.shared_experts.up_proj.output_quantizer TensorQuantizer(disabled) +model.layers.38.mlp.shared_experts.up_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.38.mlp.shared_experts.down_proj.input_quantizer TensorQuantizer(disabled) +model.layers.38.mlp.shared_experts.down_proj.output_quantizer TensorQuantizer(disabled) +model.layers.38.mlp.shared_experts.down_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.39.self_attn.q_a_proj.input_quantizer TensorQuantizer(disabled) +model.layers.39.self_attn.q_a_proj.output_quantizer TensorQuantizer(disabled) +model.layers.39.self_attn.q_a_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.39.self_attn.q_b_proj.input_quantizer TensorQuantizer(disabled) +model.layers.39.self_attn.q_b_proj.output_quantizer TensorQuantizer(disabled) +model.layers.39.self_attn.q_b_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.39.self_attn.kv_a_proj_with_mqa.input_quantizer TensorQuantizer(disabled) +model.layers.39.self_attn.kv_a_proj_with_mqa.output_quantizer TensorQuantizer(disabled) +model.layers.39.self_attn.kv_a_proj_with_mqa.weight_quantizer TensorQuantizer(disabled) +model.layers.39.self_attn.kv_b_proj.input_quantizer TensorQuantizer(disabled) +model.layers.39.self_attn.kv_b_proj.output_quantizer TensorQuantizer(disabled) +model.layers.39.self_attn.kv_b_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.39.self_attn.o_proj.input_quantizer TensorQuantizer(disabled) +model.layers.39.self_attn.o_proj.output_quantizer TensorQuantizer(disabled) +model.layers.39.self_attn.o_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.39.self_attn.q_bmm_quantizer TensorQuantizer(disabled) +model.layers.39.self_attn.k_bmm_quantizer TensorQuantizer((4, 3) bit fake per-tensor amax=4.48e+02(const) calibrator=MaxCalibrator quant) +model.layers.39.self_attn.v_bmm_quantizer TensorQuantizer((4, 3) bit fake per-tensor amax=4.48e+02(const) calibrator=MaxCalibrator quant) +model.layers.39.self_attn.p_bmm_quantizer TensorQuantizer(disabled) +model.layers.39.mlp.experts.gate_up_proj_input_quantizer TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.87e+00 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.gate_up_proj_weight_quantizers.0 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.gate_up_proj_weight_quantizers.1 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.23e-01 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.gate_up_proj_weight_quantizers.2 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.gate_up_proj_weight_quantizers.3 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.60e-01 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.gate_up_proj_weight_quantizers.4 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.29e-01 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.gate_up_proj_weight_quantizers.5 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.gate_up_proj_weight_quantizers.6 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.25e-01 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.gate_up_proj_weight_quantizers.7 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.gate_up_proj_weight_quantizers.8 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.25e-01 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.gate_up_proj_weight_quantizers.9 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.71e-01 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.gate_up_proj_weight_quantizers.10 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.gate_up_proj_weight_quantizers.11 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.gate_up_proj_weight_quantizers.12 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.77e-01 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.gate_up_proj_weight_quantizers.13 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.gate_up_proj_weight_quantizers.14 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.75e-01 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.gate_up_proj_weight_quantizers.15 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.gate_up_proj_weight_quantizers.16 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.45e-01 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.gate_up_proj_weight_quantizers.17 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.33e-01 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.gate_up_proj_weight_quantizers.18 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.53e-01 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.gate_up_proj_weight_quantizers.19 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.64e-01 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.gate_up_proj_weight_quantizers.20 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.gate_up_proj_weight_quantizers.21 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.gate_up_proj_weight_quantizers.22 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.gate_up_proj_weight_quantizers.23 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.gate_up_proj_weight_quantizers.24 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.58e-01 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.gate_up_proj_weight_quantizers.25 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.gate_up_proj_weight_quantizers.26 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.75e-01 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.gate_up_proj_weight_quantizers.27 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.15e-01 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.gate_up_proj_weight_quantizers.28 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.21e-01 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.gate_up_proj_weight_quantizers.29 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.gate_up_proj_weight_quantizers.30 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.gate_up_proj_weight_quantizers.31 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.24e-01 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.gate_up_proj_weight_quantizers.32 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.67e-01 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.gate_up_proj_weight_quantizers.33 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.26e-01 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.gate_up_proj_weight_quantizers.34 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.75e-01 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.gate_up_proj_weight_quantizers.35 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.gate_up_proj_weight_quantizers.36 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.gate_up_proj_weight_quantizers.37 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.26e-01 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.gate_up_proj_weight_quantizers.38 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.gate_up_proj_weight_quantizers.39 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=4.47e-01 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.gate_up_proj_weight_quantizers.40 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.gate_up_proj_weight_quantizers.41 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.gate_up_proj_weight_quantizers.42 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.gate_up_proj_weight_quantizers.43 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.gate_up_proj_weight_quantizers.44 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.gate_up_proj_weight_quantizers.45 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.gate_up_proj_weight_quantizers.46 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.01e-01 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.gate_up_proj_weight_quantizers.47 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.28e-01 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.gate_up_proj_weight_quantizers.48 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=5.12e-01 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.gate_up_proj_weight_quantizers.49 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=4.14e-01 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.gate_up_proj_weight_quantizers.50 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.gate_up_proj_weight_quantizers.51 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.25e-01 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.gate_up_proj_weight_quantizers.52 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.gate_up_proj_weight_quantizers.53 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.73e-01 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.gate_up_proj_weight_quantizers.54 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.gate_up_proj_weight_quantizers.55 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.gate_up_proj_weight_quantizers.56 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.gate_up_proj_weight_quantizers.57 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.36e-01 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.gate_up_proj_weight_quantizers.58 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.62e-01 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.gate_up_proj_weight_quantizers.59 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.gate_up_proj_weight_quantizers.60 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.gate_up_proj_weight_quantizers.61 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.gate_up_proj_weight_quantizers.62 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.23e-01 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.gate_up_proj_weight_quantizers.63 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=4.39e-01 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.gate_up_proj_weight_quantizers.64 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.31e-01 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.gate_up_proj_weight_quantizers.65 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.70e-01 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.gate_up_proj_weight_quantizers.66 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.25e-01 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.gate_up_proj_weight_quantizers.67 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.gate_up_proj_weight_quantizers.68 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.gate_up_proj_weight_quantizers.69 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.23e-01 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.gate_up_proj_weight_quantizers.70 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.gate_up_proj_weight_quantizers.71 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.gate_up_proj_weight_quantizers.72 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.44e-01 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.gate_up_proj_weight_quantizers.73 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.26e-01 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.gate_up_proj_weight_quantizers.74 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.gate_up_proj_weight_quantizers.75 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.46e-01 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.gate_up_proj_weight_quantizers.76 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.14e-01 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.gate_up_proj_weight_quantizers.77 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.gate_up_proj_weight_quantizers.78 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.05e-01 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.gate_up_proj_weight_quantizers.79 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.82e-01 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.gate_up_proj_weight_quantizers.80 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.14e-01 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.gate_up_proj_weight_quantizers.81 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.34e-01 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.gate_up_proj_weight_quantizers.82 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.gate_up_proj_weight_quantizers.83 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.45e-01 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.gate_up_proj_weight_quantizers.84 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.95e-01 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.gate_up_proj_weight_quantizers.85 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.gate_up_proj_weight_quantizers.86 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.25e-01 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.gate_up_proj_weight_quantizers.87 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.gate_up_proj_weight_quantizers.88 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.gate_up_proj_weight_quantizers.89 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.gate_up_proj_weight_quantizers.90 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.29e-01 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.gate_up_proj_weight_quantizers.91 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.28e-01 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.gate_up_proj_weight_quantizers.92 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.36e-01 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.gate_up_proj_weight_quantizers.93 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.gate_up_proj_weight_quantizers.94 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.gate_up_proj_weight_quantizers.95 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.gate_up_proj_weight_quantizers.96 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.gate_up_proj_weight_quantizers.97 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.77e-01 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.gate_up_proj_weight_quantizers.98 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.72e-01 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.gate_up_proj_weight_quantizers.99 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.38e-01 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.gate_up_proj_weight_quantizers.100 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.gate_up_proj_weight_quantizers.101 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.23e-01 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.gate_up_proj_weight_quantizers.102 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.55e-01 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.gate_up_proj_weight_quantizers.103 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.13e-01 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.gate_up_proj_weight_quantizers.104 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.19e-01 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.gate_up_proj_weight_quantizers.105 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.gate_up_proj_weight_quantizers.106 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.gate_up_proj_weight_quantizers.107 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.14e-01 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.gate_up_proj_weight_quantizers.108 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.gate_up_proj_weight_quantizers.109 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.16e-01 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.gate_up_proj_weight_quantizers.110 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.gate_up_proj_weight_quantizers.111 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.24e-01 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.gate_up_proj_weight_quantizers.112 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.gate_up_proj_weight_quantizers.113 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.gate_up_proj_weight_quantizers.114 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.gate_up_proj_weight_quantizers.115 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.gate_up_proj_weight_quantizers.116 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.gate_up_proj_weight_quantizers.117 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.36e-01 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.gate_up_proj_weight_quantizers.118 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.18e-01 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.gate_up_proj_weight_quantizers.119 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.21e-01 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.gate_up_proj_weight_quantizers.120 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.23e-01 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.gate_up_proj_weight_quantizers.121 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.gate_up_proj_weight_quantizers.122 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.12e-01 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.gate_up_proj_weight_quantizers.123 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.15e-01 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.gate_up_proj_weight_quantizers.124 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.17e-01 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.gate_up_proj_weight_quantizers.125 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.gate_up_proj_weight_quantizers.126 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.gate_up_proj_weight_quantizers.127 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.87e-01 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.gate_up_proj_weight_quantizers.128 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.gate_up_proj_weight_quantizers.129 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.44e-01 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.gate_up_proj_weight_quantizers.130 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.gate_up_proj_weight_quantizers.131 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.55e-01 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.gate_up_proj_weight_quantizers.132 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.gate_up_proj_weight_quantizers.133 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.gate_up_proj_weight_quantizers.134 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.gate_up_proj_weight_quantizers.135 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.gate_up_proj_weight_quantizers.136 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.gate_up_proj_weight_quantizers.137 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.83e-01 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.gate_up_proj_weight_quantizers.138 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.41e-01 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.gate_up_proj_weight_quantizers.139 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.36e-01 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.gate_up_proj_weight_quantizers.140 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.16e-01 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.gate_up_proj_weight_quantizers.141 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.99e-01 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.gate_up_proj_weight_quantizers.142 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.gate_up_proj_weight_quantizers.143 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.92e-01 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.gate_up_proj_weight_quantizers.144 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.gate_up_proj_weight_quantizers.145 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.gate_up_proj_weight_quantizers.146 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.89e-01 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.gate_up_proj_weight_quantizers.147 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.gate_up_proj_weight_quantizers.148 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.29e-01 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.gate_up_proj_weight_quantizers.149 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.gate_up_proj_weight_quantizers.150 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.33e-01 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.gate_up_proj_weight_quantizers.151 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.32e-01 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.gate_up_proj_weight_quantizers.152 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.gate_up_proj_weight_quantizers.153 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.56e-01 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.gate_up_proj_weight_quantizers.154 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.25e-01 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.gate_up_proj_weight_quantizers.155 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.76e-01 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.gate_up_proj_weight_quantizers.156 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=4.24e-01 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.gate_up_proj_weight_quantizers.157 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.gate_up_proj_weight_quantizers.158 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.18e-01 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.gate_up_proj_weight_quantizers.159 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.63e-01 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.gate_up_proj_weight_quantizers.160 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.19e-01 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.gate_up_proj_weight_quantizers.161 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.gate_up_proj_weight_quantizers.162 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.28e-01 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.gate_up_proj_weight_quantizers.163 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.gate_up_proj_weight_quantizers.164 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.gate_up_proj_weight_quantizers.165 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.gate_up_proj_weight_quantizers.166 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.29e-01 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.gate_up_proj_weight_quantizers.167 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.gate_up_proj_weight_quantizers.168 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.gate_up_proj_weight_quantizers.169 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.gate_up_proj_weight_quantizers.170 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.gate_up_proj_weight_quantizers.171 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.97e-01 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.gate_up_proj_weight_quantizers.172 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.gate_up_proj_weight_quantizers.173 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.gate_up_proj_weight_quantizers.174 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.33e-01 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.gate_up_proj_weight_quantizers.175 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.79e-01 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.gate_up_proj_weight_quantizers.176 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.gate_up_proj_weight_quantizers.177 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.gate_up_proj_weight_quantizers.178 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.gate_up_proj_weight_quantizers.179 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.gate_up_proj_weight_quantizers.180 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.83e-01 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.gate_up_proj_weight_quantizers.181 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.gate_up_proj_weight_quantizers.182 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.14e-01 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.gate_up_proj_weight_quantizers.183 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.64e-01 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.gate_up_proj_weight_quantizers.184 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.gate_up_proj_weight_quantizers.185 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.40e-01 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.gate_up_proj_weight_quantizers.186 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.gate_up_proj_weight_quantizers.187 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.gate_up_proj_weight_quantizers.188 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.gate_up_proj_weight_quantizers.189 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.24e-01 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.gate_up_proj_weight_quantizers.190 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.gate_up_proj_weight_quantizers.191 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.48e-01 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.gate_up_proj_weight_quantizers.192 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.gate_up_proj_weight_quantizers.193 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.gate_up_proj_weight_quantizers.194 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.gate_up_proj_weight_quantizers.195 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.gate_up_proj_weight_quantizers.196 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.gate_up_proj_weight_quantizers.197 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.67e-01 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.gate_up_proj_weight_quantizers.198 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.gate_up_proj_weight_quantizers.199 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.gate_up_proj_weight_quantizers.200 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.gate_up_proj_weight_quantizers.201 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.gate_up_proj_weight_quantizers.202 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.gate_up_proj_weight_quantizers.203 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.gate_up_proj_weight_quantizers.204 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.21e-01 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.gate_up_proj_weight_quantizers.205 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.gate_up_proj_weight_quantizers.206 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.gate_up_proj_weight_quantizers.207 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.gate_up_proj_weight_quantizers.208 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.15e-01 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.gate_up_proj_weight_quantizers.209 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.gate_up_proj_weight_quantizers.210 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.gate_up_proj_weight_quantizers.211 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.gate_up_proj_weight_quantizers.212 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.91e-01 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.gate_up_proj_weight_quantizers.213 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.22e-01 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.gate_up_proj_weight_quantizers.214 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.gate_up_proj_weight_quantizers.215 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.53e-01 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.gate_up_proj_weight_quantizers.216 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.gate_up_proj_weight_quantizers.217 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.gate_up_proj_weight_quantizers.218 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.gate_up_proj_weight_quantizers.219 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.gate_up_proj_weight_quantizers.220 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.44e-01 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.gate_up_proj_weight_quantizers.221 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.00e-01 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.gate_up_proj_weight_quantizers.222 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.gate_up_proj_weight_quantizers.223 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.gate_up_proj_weight_quantizers.224 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.00e-01 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.gate_up_proj_weight_quantizers.225 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.gate_up_proj_weight_quantizers.226 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.gate_up_proj_weight_quantizers.227 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.gate_up_proj_weight_quantizers.228 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.49e-01 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.gate_up_proj_weight_quantizers.229 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.83e-01 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.gate_up_proj_weight_quantizers.230 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.gate_up_proj_weight_quantizers.231 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=6.45e-01 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.gate_up_proj_weight_quantizers.232 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.gate_up_proj_weight_quantizers.233 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.gate_up_proj_weight_quantizers.234 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.gate_up_proj_weight_quantizers.235 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.49e-01 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.gate_up_proj_weight_quantizers.236 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.gate_up_proj_weight_quantizers.237 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.64e-01 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.gate_up_proj_weight_quantizers.238 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.63e-01 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.gate_up_proj_weight_quantizers.239 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.31e-01 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.gate_up_proj_weight_quantizers.240 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.gate_up_proj_weight_quantizers.241 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.75e-01 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.gate_up_proj_weight_quantizers.242 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.gate_up_proj_weight_quantizers.243 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.gate_up_proj_weight_quantizers.244 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.gate_up_proj_weight_quantizers.245 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.45e-01 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.gate_up_proj_weight_quantizers.246 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.86e-01 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.gate_up_proj_weight_quantizers.247 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.gate_up_proj_weight_quantizers.248 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.gate_up_proj_weight_quantizers.249 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.gate_up_proj_weight_quantizers.250 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.gate_up_proj_weight_quantizers.251 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.gate_up_proj_weight_quantizers.252 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.gate_up_proj_weight_quantizers.253 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.gate_up_proj_weight_quantizers.254 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.36e-01 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.gate_up_proj_weight_quantizers.255 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.down_proj_input_quantizer TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.57e+01 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.down_proj_weight_quantizers.0 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.down_proj_weight_quantizers.1 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.down_proj_weight_quantizers.2 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.down_proj_weight_quantizers.3 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.97e-01 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.down_proj_weight_quantizers.4 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.down_proj_weight_quantizers.5 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.down_proj_weight_quantizers.6 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.down_proj_weight_quantizers.7 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.down_proj_weight_quantizers.8 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.down_proj_weight_quantizers.9 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.74e-01 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.down_proj_weight_quantizers.10 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.down_proj_weight_quantizers.11 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.down_proj_weight_quantizers.12 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.40e-02 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.down_proj_weight_quantizers.13 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.down_proj_weight_quantizers.14 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.down_proj_weight_quantizers.15 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.down_proj_weight_quantizers.16 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.down_proj_weight_quantizers.17 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.23e-01 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.down_proj_weight_quantizers.18 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.18e-01 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.down_proj_weight_quantizers.19 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.down_proj_weight_quantizers.20 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.down_proj_weight_quantizers.21 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.03e-01 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.down_proj_weight_quantizers.22 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.down_proj_weight_quantizers.23 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.down_proj_weight_quantizers.24 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.down_proj_weight_quantizers.25 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.down_proj_weight_quantizers.26 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.down_proj_weight_quantizers.27 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.down_proj_weight_quantizers.28 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.23e-01 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.down_proj_weight_quantizers.29 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.down_proj_weight_quantizers.30 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.down_proj_weight_quantizers.31 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.down_proj_weight_quantizers.32 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.down_proj_weight_quantizers.33 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.down_proj_weight_quantizers.34 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.43e-01 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.down_proj_weight_quantizers.35 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.down_proj_weight_quantizers.36 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.down_proj_weight_quantizers.37 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.60e-01 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.down_proj_weight_quantizers.38 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.down_proj_weight_quantizers.39 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.down_proj_weight_quantizers.40 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.down_proj_weight_quantizers.41 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.down_proj_weight_quantizers.42 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.down_proj_weight_quantizers.43 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.down_proj_weight_quantizers.44 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.down_proj_weight_quantizers.45 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.down_proj_weight_quantizers.46 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.37e-01 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.down_proj_weight_quantizers.47 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.91e-01 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.down_proj_weight_quantizers.48 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.68e-01 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.down_proj_weight_quantizers.49 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.97e-01 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.down_proj_weight_quantizers.50 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.59e-01 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.down_proj_weight_quantizers.51 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.42e-01 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.down_proj_weight_quantizers.52 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.43e-01 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.down_proj_weight_quantizers.53 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.down_proj_weight_quantizers.54 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.down_proj_weight_quantizers.55 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.down_proj_weight_quantizers.56 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.down_proj_weight_quantizers.57 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.down_proj_weight_quantizers.58 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.down_proj_weight_quantizers.59 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.down_proj_weight_quantizers.60 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.59e-02 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.down_proj_weight_quantizers.61 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.down_proj_weight_quantizers.62 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.59e-01 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.down_proj_weight_quantizers.63 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.73e-01 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.down_proj_weight_quantizers.64 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.16e-01 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.down_proj_weight_quantizers.65 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.down_proj_weight_quantizers.66 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.down_proj_weight_quantizers.67 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.down_proj_weight_quantizers.68 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.down_proj_weight_quantizers.69 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.45e-01 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.down_proj_weight_quantizers.70 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.down_proj_weight_quantizers.71 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.down_proj_weight_quantizers.72 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.44e-01 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.down_proj_weight_quantizers.73 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.37e-01 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.down_proj_weight_quantizers.74 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.down_proj_weight_quantizers.75 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.22e-01 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.down_proj_weight_quantizers.76 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.down_proj_weight_quantizers.77 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.down_proj_weight_quantizers.78 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.down_proj_weight_quantizers.79 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.32e-01 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.down_proj_weight_quantizers.80 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.down_proj_weight_quantizers.81 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.64e-02 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.down_proj_weight_quantizers.82 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.down_proj_weight_quantizers.83 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.56e-01 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.down_proj_weight_quantizers.84 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.27e-01 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.down_proj_weight_quantizers.85 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.down_proj_weight_quantizers.86 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.19e-01 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.down_proj_weight_quantizers.87 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.down_proj_weight_quantizers.88 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.59e-01 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.down_proj_weight_quantizers.89 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.14e-01 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.down_proj_weight_quantizers.90 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.down_proj_weight_quantizers.91 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.down_proj_weight_quantizers.92 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.down_proj_weight_quantizers.93 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.down_proj_weight_quantizers.94 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.down_proj_weight_quantizers.95 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.down_proj_weight_quantizers.96 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.down_proj_weight_quantizers.97 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.14e-01 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.down_proj_weight_quantizers.98 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.80e-01 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.down_proj_weight_quantizers.99 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.23e-01 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.down_proj_weight_quantizers.100 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.down_proj_weight_quantizers.101 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.down_proj_weight_quantizers.102 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.47e-01 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.down_proj_weight_quantizers.103 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.down_proj_weight_quantizers.104 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.down_proj_weight_quantizers.105 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.64e-02 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.down_proj_weight_quantizers.106 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.59e-02 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.down_proj_weight_quantizers.107 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.59e-02 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.down_proj_weight_quantizers.108 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.down_proj_weight_quantizers.109 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.down_proj_weight_quantizers.110 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.down_proj_weight_quantizers.111 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.down_proj_weight_quantizers.112 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.down_proj_weight_quantizers.113 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.down_proj_weight_quantizers.114 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.down_proj_weight_quantizers.115 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.down_proj_weight_quantizers.116 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.down_proj_weight_quantizers.117 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.down_proj_weight_quantizers.118 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.down_proj_weight_quantizers.119 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.54e-02 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.down_proj_weight_quantizers.120 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.down_proj_weight_quantizers.121 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.down_proj_weight_quantizers.122 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.down_proj_weight_quantizers.123 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.77e-01 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.down_proj_weight_quantizers.124 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.down_proj_weight_quantizers.125 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.down_proj_weight_quantizers.126 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.down_proj_weight_quantizers.127 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.77e-01 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.down_proj_weight_quantizers.128 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.down_proj_weight_quantizers.129 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.54e-02 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.down_proj_weight_quantizers.130 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.down_proj_weight_quantizers.131 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.13e-01 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.down_proj_weight_quantizers.132 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.down_proj_weight_quantizers.133 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.down_proj_weight_quantizers.134 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.54e-02 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.down_proj_weight_quantizers.135 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.down_proj_weight_quantizers.136 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.down_proj_weight_quantizers.137 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.down_proj_weight_quantizers.138 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.down_proj_weight_quantizers.139 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.down_proj_weight_quantizers.140 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.down_proj_weight_quantizers.141 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.down_proj_weight_quantizers.142 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.down_proj_weight_quantizers.143 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.29e-01 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.down_proj_weight_quantizers.144 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.down_proj_weight_quantizers.145 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.down_proj_weight_quantizers.146 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.77e-01 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.down_proj_weight_quantizers.147 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.down_proj_weight_quantizers.148 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.91e-01 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.down_proj_weight_quantizers.149 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.down_proj_weight_quantizers.150 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.81e-01 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.down_proj_weight_quantizers.151 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.down_proj_weight_quantizers.152 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.down_proj_weight_quantizers.153 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.down_proj_weight_quantizers.154 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.down_proj_weight_quantizers.155 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.38e-01 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.down_proj_weight_quantizers.156 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.36e-01 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.down_proj_weight_quantizers.157 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.down_proj_weight_quantizers.158 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.down_proj_weight_quantizers.159 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.down_proj_weight_quantizers.160 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.down_proj_weight_quantizers.161 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.20e-02 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.down_proj_weight_quantizers.162 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.68e-01 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.down_proj_weight_quantizers.163 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.down_proj_weight_quantizers.164 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.down_proj_weight_quantizers.165 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.down_proj_weight_quantizers.166 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.21e-01 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.down_proj_weight_quantizers.167 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.down_proj_weight_quantizers.168 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.down_proj_weight_quantizers.169 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.down_proj_weight_quantizers.170 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.down_proj_weight_quantizers.171 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.24e-01 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.down_proj_weight_quantizers.172 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.down_proj_weight_quantizers.173 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.down_proj_weight_quantizers.174 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.13e-01 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.down_proj_weight_quantizers.175 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.18e-01 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.down_proj_weight_quantizers.176 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.59e-02 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.down_proj_weight_quantizers.177 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.down_proj_weight_quantizers.178 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.down_proj_weight_quantizers.179 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.down_proj_weight_quantizers.180 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.down_proj_weight_quantizers.181 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.down_proj_weight_quantizers.182 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.down_proj_weight_quantizers.183 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.57e-01 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.down_proj_weight_quantizers.184 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.down_proj_weight_quantizers.185 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.down_proj_weight_quantizers.186 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.54e-02 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.down_proj_weight_quantizers.187 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.down_proj_weight_quantizers.188 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.down_proj_weight_quantizers.189 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.down_proj_weight_quantizers.190 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.down_proj_weight_quantizers.191 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.32e-01 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.down_proj_weight_quantizers.192 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.down_proj_weight_quantizers.193 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.down_proj_weight_quantizers.194 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.down_proj_weight_quantizers.195 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.down_proj_weight_quantizers.196 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.down_proj_weight_quantizers.197 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.down_proj_weight_quantizers.198 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.down_proj_weight_quantizers.199 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.down_proj_weight_quantizers.200 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.31e-01 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.down_proj_weight_quantizers.201 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.down_proj_weight_quantizers.202 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.down_proj_weight_quantizers.203 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.down_proj_weight_quantizers.204 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.down_proj_weight_quantizers.205 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.down_proj_weight_quantizers.206 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.down_proj_weight_quantizers.207 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.down_proj_weight_quantizers.208 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.40e-02 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.down_proj_weight_quantizers.209 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.59e-02 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.down_proj_weight_quantizers.210 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.down_proj_weight_quantizers.211 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.down_proj_weight_quantizers.212 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.25e-01 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.down_proj_weight_quantizers.213 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.30e-02 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.down_proj_weight_quantizers.214 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.down_proj_weight_quantizers.215 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.down_proj_weight_quantizers.216 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.down_proj_weight_quantizers.217 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.down_proj_weight_quantizers.218 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.down_proj_weight_quantizers.219 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.down_proj_weight_quantizers.220 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.13e-01 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.down_proj_weight_quantizers.221 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.35e-01 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.down_proj_weight_quantizers.222 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.down_proj_weight_quantizers.223 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.down_proj_weight_quantizers.224 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.down_proj_weight_quantizers.225 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.down_proj_weight_quantizers.226 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.down_proj_weight_quantizers.227 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.45e-02 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.down_proj_weight_quantizers.228 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.down_proj_weight_quantizers.229 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.64e-02 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.down_proj_weight_quantizers.230 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.28e-01 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.down_proj_weight_quantizers.231 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.72e-01 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.down_proj_weight_quantizers.232 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.24e-01 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.down_proj_weight_quantizers.233 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.down_proj_weight_quantizers.234 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.down_proj_weight_quantizers.235 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.52e-01 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.down_proj_weight_quantizers.236 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.down_proj_weight_quantizers.237 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.down_proj_weight_quantizers.238 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.46e-01 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.down_proj_weight_quantizers.239 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.down_proj_weight_quantizers.240 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.down_proj_weight_quantizers.241 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.down_proj_weight_quantizers.242 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.59e-02 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.down_proj_weight_quantizers.243 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.down_proj_weight_quantizers.244 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.46e-01 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.down_proj_weight_quantizers.245 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.down_proj_weight_quantizers.246 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.20e-01 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.down_proj_weight_quantizers.247 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.down_proj_weight_quantizers.248 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.down_proj_weight_quantizers.249 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.down_proj_weight_quantizers.250 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.50e-02 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.down_proj_weight_quantizers.251 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.down_proj_weight_quantizers.252 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.down_proj_weight_quantizers.253 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.88e-01 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.down_proj_weight_quantizers.254 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.24e-01 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.down_proj_weight_quantizers.255 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.39.mlp.shared_experts.gate_proj.input_quantizer TensorQuantizer(disabled) +model.layers.39.mlp.shared_experts.gate_proj.output_quantizer TensorQuantizer(disabled) +model.layers.39.mlp.shared_experts.gate_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.39.mlp.shared_experts.up_proj.input_quantizer TensorQuantizer(disabled) +model.layers.39.mlp.shared_experts.up_proj.output_quantizer TensorQuantizer(disabled) +model.layers.39.mlp.shared_experts.up_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.39.mlp.shared_experts.down_proj.input_quantizer TensorQuantizer(disabled) +model.layers.39.mlp.shared_experts.down_proj.output_quantizer TensorQuantizer(disabled) +model.layers.39.mlp.shared_experts.down_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.40.self_attn.q_a_proj.input_quantizer TensorQuantizer(disabled) +model.layers.40.self_attn.q_a_proj.output_quantizer TensorQuantizer(disabled) +model.layers.40.self_attn.q_a_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.40.self_attn.q_b_proj.input_quantizer TensorQuantizer(disabled) +model.layers.40.self_attn.q_b_proj.output_quantizer TensorQuantizer(disabled) +model.layers.40.self_attn.q_b_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.40.self_attn.kv_a_proj_with_mqa.input_quantizer TensorQuantizer(disabled) +model.layers.40.self_attn.kv_a_proj_with_mqa.output_quantizer TensorQuantizer(disabled) +model.layers.40.self_attn.kv_a_proj_with_mqa.weight_quantizer TensorQuantizer(disabled) +model.layers.40.self_attn.kv_b_proj.input_quantizer TensorQuantizer(disabled) +model.layers.40.self_attn.kv_b_proj.output_quantizer TensorQuantizer(disabled) +model.layers.40.self_attn.kv_b_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.40.self_attn.o_proj.input_quantizer TensorQuantizer(disabled) +model.layers.40.self_attn.o_proj.output_quantizer TensorQuantizer(disabled) +model.layers.40.self_attn.o_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.40.self_attn.q_bmm_quantizer TensorQuantizer(disabled) +model.layers.40.self_attn.k_bmm_quantizer TensorQuantizer((4, 3) bit fake per-tensor amax=4.48e+02(const) calibrator=MaxCalibrator quant) +model.layers.40.self_attn.v_bmm_quantizer TensorQuantizer((4, 3) bit fake per-tensor amax=4.48e+02(const) calibrator=MaxCalibrator quant) +model.layers.40.self_attn.p_bmm_quantizer TensorQuantizer(disabled) +model.layers.40.mlp.experts.gate_up_proj_input_quantizer TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.08e+00 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.gate_up_proj_weight_quantizers.0 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.86e-01 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.gate_up_proj_weight_quantizers.1 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.gate_up_proj_weight_quantizers.2 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.16e-01 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.gate_up_proj_weight_quantizers.3 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.18e-01 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.gate_up_proj_weight_quantizers.4 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.40e-01 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.gate_up_proj_weight_quantizers.5 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.20e-01 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.gate_up_proj_weight_quantizers.6 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.21e-01 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.gate_up_proj_weight_quantizers.7 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.66e-01 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.gate_up_proj_weight_quantizers.8 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.gate_up_proj_weight_quantizers.9 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.gate_up_proj_weight_quantizers.10 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.gate_up_proj_weight_quantizers.11 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.gate_up_proj_weight_quantizers.12 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.20e-01 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.gate_up_proj_weight_quantizers.13 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.gate_up_proj_weight_quantizers.14 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.98e-01 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.gate_up_proj_weight_quantizers.15 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.91e-01 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.gate_up_proj_weight_quantizers.16 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.gate_up_proj_weight_quantizers.17 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.gate_up_proj_weight_quantizers.18 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.gate_up_proj_weight_quantizers.19 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.46e-01 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.gate_up_proj_weight_quantizers.20 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.gate_up_proj_weight_quantizers.21 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.gate_up_proj_weight_quantizers.22 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.28e-01 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.gate_up_proj_weight_quantizers.23 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.gate_up_proj_weight_quantizers.24 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.gate_up_proj_weight_quantizers.25 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.18e-01 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.gate_up_proj_weight_quantizers.26 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.gate_up_proj_weight_quantizers.27 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.41e-01 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.gate_up_proj_weight_quantizers.28 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.59e-01 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.gate_up_proj_weight_quantizers.29 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.gate_up_proj_weight_quantizers.30 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.gate_up_proj_weight_quantizers.31 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.15e-01 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.gate_up_proj_weight_quantizers.32 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.gate_up_proj_weight_quantizers.33 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.28e-01 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.gate_up_proj_weight_quantizers.34 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.24e-01 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.gate_up_proj_weight_quantizers.35 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.59e-01 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.gate_up_proj_weight_quantizers.36 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.23e-01 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.gate_up_proj_weight_quantizers.37 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.gate_up_proj_weight_quantizers.38 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.21e-01 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.gate_up_proj_weight_quantizers.39 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.09e-01 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.gate_up_proj_weight_quantizers.40 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.22e-01 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.gate_up_proj_weight_quantizers.41 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.gate_up_proj_weight_quantizers.42 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.gate_up_proj_weight_quantizers.43 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.gate_up_proj_weight_quantizers.44 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.01e-01 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.gate_up_proj_weight_quantizers.45 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.gate_up_proj_weight_quantizers.46 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.gate_up_proj_weight_quantizers.47 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.gate_up_proj_weight_quantizers.48 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.gate_up_proj_weight_quantizers.49 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.gate_up_proj_weight_quantizers.50 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.44e-01 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.gate_up_proj_weight_quantizers.51 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.gate_up_proj_weight_quantizers.52 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.20e-01 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.gate_up_proj_weight_quantizers.53 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.gate_up_proj_weight_quantizers.54 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.33e-01 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.gate_up_proj_weight_quantizers.55 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.gate_up_proj_weight_quantizers.56 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.gate_up_proj_weight_quantizers.57 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.gate_up_proj_weight_quantizers.58 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.45e-01 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.gate_up_proj_weight_quantizers.59 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.gate_up_proj_weight_quantizers.60 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.gate_up_proj_weight_quantizers.61 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.gate_up_proj_weight_quantizers.62 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.37e-01 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.gate_up_proj_weight_quantizers.63 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.45e-01 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.gate_up_proj_weight_quantizers.64 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.gate_up_proj_weight_quantizers.65 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.gate_up_proj_weight_quantizers.66 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.gate_up_proj_weight_quantizers.67 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.30e-01 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.gate_up_proj_weight_quantizers.68 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.gate_up_proj_weight_quantizers.69 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.24e-01 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.gate_up_proj_weight_quantizers.70 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.gate_up_proj_weight_quantizers.71 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.gate_up_proj_weight_quantizers.72 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.gate_up_proj_weight_quantizers.73 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.95e-01 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.gate_up_proj_weight_quantizers.74 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.gate_up_proj_weight_quantizers.75 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.26e-01 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.gate_up_proj_weight_quantizers.76 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.gate_up_proj_weight_quantizers.77 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.gate_up_proj_weight_quantizers.78 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.54e-01 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.gate_up_proj_weight_quantizers.79 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.gate_up_proj_weight_quantizers.80 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.28e-01 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.gate_up_proj_weight_quantizers.81 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.36e-01 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.gate_up_proj_weight_quantizers.82 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.32e-01 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.gate_up_proj_weight_quantizers.83 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.24e-01 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.gate_up_proj_weight_quantizers.84 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.95e-01 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.gate_up_proj_weight_quantizers.85 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.93e-01 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.gate_up_proj_weight_quantizers.86 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.gate_up_proj_weight_quantizers.87 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.24e-01 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.gate_up_proj_weight_quantizers.88 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.gate_up_proj_weight_quantizers.89 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.gate_up_proj_weight_quantizers.90 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.gate_up_proj_weight_quantizers.91 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.63e-01 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.gate_up_proj_weight_quantizers.92 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.45e-01 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.gate_up_proj_weight_quantizers.93 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.gate_up_proj_weight_quantizers.94 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.gate_up_proj_weight_quantizers.95 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.13e-01 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.gate_up_proj_weight_quantizers.96 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=5.04e-01 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.gate_up_proj_weight_quantizers.97 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.gate_up_proj_weight_quantizers.98 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.52e-01 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.gate_up_proj_weight_quantizers.99 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.16e-01 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.gate_up_proj_weight_quantizers.100 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.gate_up_proj_weight_quantizers.101 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.gate_up_proj_weight_quantizers.102 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.gate_up_proj_weight_quantizers.103 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.61e-01 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.gate_up_proj_weight_quantizers.104 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.34e-01 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.gate_up_proj_weight_quantizers.105 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.gate_up_proj_weight_quantizers.106 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.gate_up_proj_weight_quantizers.107 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.gate_up_proj_weight_quantizers.108 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.gate_up_proj_weight_quantizers.109 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.gate_up_proj_weight_quantizers.110 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.61e-01 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.gate_up_proj_weight_quantizers.111 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.gate_up_proj_weight_quantizers.112 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.gate_up_proj_weight_quantizers.113 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.35e-01 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.gate_up_proj_weight_quantizers.114 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.gate_up_proj_weight_quantizers.115 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.43e-01 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.gate_up_proj_weight_quantizers.116 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.30e-01 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.gate_up_proj_weight_quantizers.117 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.13e-01 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.gate_up_proj_weight_quantizers.118 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.gate_up_proj_weight_quantizers.119 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.gate_up_proj_weight_quantizers.120 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.gate_up_proj_weight_quantizers.121 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.13e-01 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.gate_up_proj_weight_quantizers.122 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.69e-01 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.gate_up_proj_weight_quantizers.123 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.65e-01 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.gate_up_proj_weight_quantizers.124 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.83e-01 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.gate_up_proj_weight_quantizers.125 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.gate_up_proj_weight_quantizers.126 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.gate_up_proj_weight_quantizers.127 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.46e-01 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.gate_up_proj_weight_quantizers.128 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.gate_up_proj_weight_quantizers.129 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.85e-01 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.gate_up_proj_weight_quantizers.130 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.39e-01 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.gate_up_proj_weight_quantizers.131 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.75e-01 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.gate_up_proj_weight_quantizers.132 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.21e-01 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.gate_up_proj_weight_quantizers.133 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.gate_up_proj_weight_quantizers.134 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.33e-01 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.gate_up_proj_weight_quantizers.135 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.gate_up_proj_weight_quantizers.136 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.60e-01 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.gate_up_proj_weight_quantizers.137 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.67e-01 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.gate_up_proj_weight_quantizers.138 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.gate_up_proj_weight_quantizers.139 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.33e-01 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.gate_up_proj_weight_quantizers.140 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.04e-01 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.gate_up_proj_weight_quantizers.141 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.gate_up_proj_weight_quantizers.142 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.gate_up_proj_weight_quantizers.143 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.16e-01 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.gate_up_proj_weight_quantizers.144 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.46e-01 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.gate_up_proj_weight_quantizers.145 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.gate_up_proj_weight_quantizers.146 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.23e-01 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.gate_up_proj_weight_quantizers.147 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.gate_up_proj_weight_quantizers.148 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.19e-01 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.gate_up_proj_weight_quantizers.149 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.14e-01 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.gate_up_proj_weight_quantizers.150 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.25e-01 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.gate_up_proj_weight_quantizers.151 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=4.24e-01 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.gate_up_proj_weight_quantizers.152 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.38e-01 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.gate_up_proj_weight_quantizers.153 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.gate_up_proj_weight_quantizers.154 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.gate_up_proj_weight_quantizers.155 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.21e-01 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.gate_up_proj_weight_quantizers.156 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.gate_up_proj_weight_quantizers.157 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.79e-01 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.gate_up_proj_weight_quantizers.158 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.22e-01 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.gate_up_proj_weight_quantizers.159 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.gate_up_proj_weight_quantizers.160 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.30e-01 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.gate_up_proj_weight_quantizers.161 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=4.10e-01 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.gate_up_proj_weight_quantizers.162 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.gate_up_proj_weight_quantizers.163 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.gate_up_proj_weight_quantizers.164 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.gate_up_proj_weight_quantizers.165 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.78e-01 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.gate_up_proj_weight_quantizers.166 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.25e-01 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.gate_up_proj_weight_quantizers.167 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.gate_up_proj_weight_quantizers.168 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.gate_up_proj_weight_quantizers.169 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.gate_up_proj_weight_quantizers.170 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.25e-01 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.gate_up_proj_weight_quantizers.171 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.gate_up_proj_weight_quantizers.172 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.47e-01 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.gate_up_proj_weight_quantizers.173 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.32e-01 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.gate_up_proj_weight_quantizers.174 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=5.23e-01 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.gate_up_proj_weight_quantizers.175 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.gate_up_proj_weight_quantizers.176 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.gate_up_proj_weight_quantizers.177 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.gate_up_proj_weight_quantizers.178 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.gate_up_proj_weight_quantizers.179 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.93e-01 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.gate_up_proj_weight_quantizers.180 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.15e-01 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.gate_up_proj_weight_quantizers.181 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.69e-01 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.gate_up_proj_weight_quantizers.182 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.58e-01 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.gate_up_proj_weight_quantizers.183 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.17e-01 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.gate_up_proj_weight_quantizers.184 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.14e-01 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.gate_up_proj_weight_quantizers.185 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.gate_up_proj_weight_quantizers.186 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.22e-01 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.gate_up_proj_weight_quantizers.187 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=4.57e-01 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.gate_up_proj_weight_quantizers.188 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.gate_up_proj_weight_quantizers.189 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.gate_up_proj_weight_quantizers.190 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.22e-01 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.gate_up_proj_weight_quantizers.191 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.gate_up_proj_weight_quantizers.192 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.gate_up_proj_weight_quantizers.193 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.59e-01 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.gate_up_proj_weight_quantizers.194 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.gate_up_proj_weight_quantizers.195 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.gate_up_proj_weight_quantizers.196 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.69e-01 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.gate_up_proj_weight_quantizers.197 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.gate_up_proj_weight_quantizers.198 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.gate_up_proj_weight_quantizers.199 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.35e-01 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.gate_up_proj_weight_quantizers.200 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.24e-01 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.gate_up_proj_weight_quantizers.201 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.gate_up_proj_weight_quantizers.202 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.gate_up_proj_weight_quantizers.203 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.75e-01 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.gate_up_proj_weight_quantizers.204 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.gate_up_proj_weight_quantizers.205 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.gate_up_proj_weight_quantizers.206 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.22e-01 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.gate_up_proj_weight_quantizers.207 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.gate_up_proj_weight_quantizers.208 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.35e-01 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.gate_up_proj_weight_quantizers.209 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.55e-01 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.gate_up_proj_weight_quantizers.210 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.11e-01 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.gate_up_proj_weight_quantizers.211 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.gate_up_proj_weight_quantizers.212 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.gate_up_proj_weight_quantizers.213 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.70e-01 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.gate_up_proj_weight_quantizers.214 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.gate_up_proj_weight_quantizers.215 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.19e-01 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.gate_up_proj_weight_quantizers.216 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.13e-01 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.gate_up_proj_weight_quantizers.217 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.gate_up_proj_weight_quantizers.218 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.65e-01 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.gate_up_proj_weight_quantizers.219 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.gate_up_proj_weight_quantizers.220 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.gate_up_proj_weight_quantizers.221 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.gate_up_proj_weight_quantizers.222 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.61e-01 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.gate_up_proj_weight_quantizers.223 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.gate_up_proj_weight_quantizers.224 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.gate_up_proj_weight_quantizers.225 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.46e-01 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.gate_up_proj_weight_quantizers.226 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.15e-01 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.gate_up_proj_weight_quantizers.227 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.91e-01 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.gate_up_proj_weight_quantizers.228 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.31e-01 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.gate_up_proj_weight_quantizers.229 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.gate_up_proj_weight_quantizers.230 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.17e-01 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.gate_up_proj_weight_quantizers.231 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.gate_up_proj_weight_quantizers.232 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.gate_up_proj_weight_quantizers.233 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.gate_up_proj_weight_quantizers.234 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.29e-01 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.gate_up_proj_weight_quantizers.235 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.25e-01 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.gate_up_proj_weight_quantizers.236 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.gate_up_proj_weight_quantizers.237 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.44e-01 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.gate_up_proj_weight_quantizers.238 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.78e-01 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.gate_up_proj_weight_quantizers.239 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.24e-01 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.gate_up_proj_weight_quantizers.240 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.29e-01 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.gate_up_proj_weight_quantizers.241 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.gate_up_proj_weight_quantizers.242 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.31e-01 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.gate_up_proj_weight_quantizers.243 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.94e-01 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.gate_up_proj_weight_quantizers.244 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.gate_up_proj_weight_quantizers.245 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.69e-01 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.gate_up_proj_weight_quantizers.246 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.gate_up_proj_weight_quantizers.247 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.17e-01 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.gate_up_proj_weight_quantizers.248 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.23e-01 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.gate_up_proj_weight_quantizers.249 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=6.17e-01 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.gate_up_proj_weight_quantizers.250 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.gate_up_proj_weight_quantizers.251 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.gate_up_proj_weight_quantizers.252 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.97e-01 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.gate_up_proj_weight_quantizers.253 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.gate_up_proj_weight_quantizers.254 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.gate_up_proj_weight_quantizers.255 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.down_proj_input_quantizer TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.82e+01 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.down_proj_weight_quantizers.0 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.down_proj_weight_quantizers.1 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.down_proj_weight_quantizers.2 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.down_proj_weight_quantizers.3 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.59e-02 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.down_proj_weight_quantizers.4 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.down_proj_weight_quantizers.5 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.down_proj_weight_quantizers.6 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.25e-01 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.down_proj_weight_quantizers.7 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.28e-01 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.down_proj_weight_quantizers.8 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.down_proj_weight_quantizers.9 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.down_proj_weight_quantizers.10 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.down_proj_weight_quantizers.11 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.64e-02 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.down_proj_weight_quantizers.12 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.down_proj_weight_quantizers.13 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.down_proj_weight_quantizers.14 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.down_proj_weight_quantizers.15 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.down_proj_weight_quantizers.16 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.down_proj_weight_quantizers.17 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.down_proj_weight_quantizers.18 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.64e-02 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.down_proj_weight_quantizers.19 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.down_proj_weight_quantizers.20 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.down_proj_weight_quantizers.21 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.down_proj_weight_quantizers.22 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.down_proj_weight_quantizers.23 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.down_proj_weight_quantizers.24 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.down_proj_weight_quantizers.25 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.down_proj_weight_quantizers.26 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.down_proj_weight_quantizers.27 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.down_proj_weight_quantizers.28 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.down_proj_weight_quantizers.29 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.23e-01 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.down_proj_weight_quantizers.30 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.down_proj_weight_quantizers.31 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.down_proj_weight_quantizers.32 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.down_proj_weight_quantizers.33 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.down_proj_weight_quantizers.34 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.97e-01 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.down_proj_weight_quantizers.35 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.down_proj_weight_quantizers.36 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.down_proj_weight_quantizers.37 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.down_proj_weight_quantizers.38 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.down_proj_weight_quantizers.39 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.down_proj_weight_quantizers.40 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.down_proj_weight_quantizers.41 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.down_proj_weight_quantizers.42 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.down_proj_weight_quantizers.43 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.35e-02 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.down_proj_weight_quantizers.44 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.46e-01 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.down_proj_weight_quantizers.45 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.down_proj_weight_quantizers.46 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.down_proj_weight_quantizers.47 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.down_proj_weight_quantizers.48 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.down_proj_weight_quantizers.49 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.down_proj_weight_quantizers.50 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.93e-01 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.down_proj_weight_quantizers.51 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.down_proj_weight_quantizers.52 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.down_proj_weight_quantizers.53 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.down_proj_weight_quantizers.54 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.58e-01 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.down_proj_weight_quantizers.55 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.down_proj_weight_quantizers.56 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.down_proj_weight_quantizers.57 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.59e-02 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.down_proj_weight_quantizers.58 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.down_proj_weight_quantizers.59 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.down_proj_weight_quantizers.60 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.down_proj_weight_quantizers.61 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.down_proj_weight_quantizers.62 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.37e-01 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.down_proj_weight_quantizers.63 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.25e-02 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.down_proj_weight_quantizers.64 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.down_proj_weight_quantizers.65 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.down_proj_weight_quantizers.66 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.down_proj_weight_quantizers.67 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.down_proj_weight_quantizers.68 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.54e-02 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.down_proj_weight_quantizers.69 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.down_proj_weight_quantizers.70 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.down_proj_weight_quantizers.71 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.down_proj_weight_quantizers.72 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.down_proj_weight_quantizers.73 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.down_proj_weight_quantizers.74 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.down_proj_weight_quantizers.75 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.63e-01 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.down_proj_weight_quantizers.76 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.54e-02 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.down_proj_weight_quantizers.77 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.down_proj_weight_quantizers.78 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.50e-01 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.down_proj_weight_quantizers.79 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.down_proj_weight_quantizers.80 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.down_proj_weight_quantizers.81 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.down_proj_weight_quantizers.82 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.97e-01 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.down_proj_weight_quantizers.83 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.down_proj_weight_quantizers.84 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.40e-01 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.down_proj_weight_quantizers.85 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.down_proj_weight_quantizers.86 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.down_proj_weight_quantizers.87 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.down_proj_weight_quantizers.88 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.down_proj_weight_quantizers.89 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.69e-01 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.down_proj_weight_quantizers.90 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.down_proj_weight_quantizers.91 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.down_proj_weight_quantizers.92 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.down_proj_weight_quantizers.93 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.down_proj_weight_quantizers.94 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.64e-02 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.down_proj_weight_quantizers.95 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.down_proj_weight_quantizers.96 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.18e-01 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.down_proj_weight_quantizers.97 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.down_proj_weight_quantizers.98 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.down_proj_weight_quantizers.99 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.15e-01 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.down_proj_weight_quantizers.100 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.down_proj_weight_quantizers.101 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.down_proj_weight_quantizers.102 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.down_proj_weight_quantizers.103 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.23e-01 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.down_proj_weight_quantizers.104 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.down_proj_weight_quantizers.105 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.45e-02 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.down_proj_weight_quantizers.106 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.down_proj_weight_quantizers.107 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.down_proj_weight_quantizers.108 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.down_proj_weight_quantizers.109 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.down_proj_weight_quantizers.110 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.down_proj_weight_quantizers.111 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.down_proj_weight_quantizers.112 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.down_proj_weight_quantizers.113 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.down_proj_weight_quantizers.114 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.down_proj_weight_quantizers.115 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.down_proj_weight_quantizers.116 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.down_proj_weight_quantizers.117 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.down_proj_weight_quantizers.118 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.down_proj_weight_quantizers.119 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.down_proj_weight_quantizers.120 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.down_proj_weight_quantizers.121 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.down_proj_weight_quantizers.122 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.45e-01 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.down_proj_weight_quantizers.123 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.down_proj_weight_quantizers.124 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.34e-01 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.down_proj_weight_quantizers.125 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.down_proj_weight_quantizers.126 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.down_proj_weight_quantizers.127 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.down_proj_weight_quantizers.128 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.54e-02 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.down_proj_weight_quantizers.129 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.66e-01 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.down_proj_weight_quantizers.130 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.down_proj_weight_quantizers.131 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.down_proj_weight_quantizers.132 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.down_proj_weight_quantizers.133 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.down_proj_weight_quantizers.134 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.down_proj_weight_quantizers.135 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.down_proj_weight_quantizers.136 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.79e-01 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.down_proj_weight_quantizers.137 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.down_proj_weight_quantizers.138 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.down_proj_weight_quantizers.139 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.down_proj_weight_quantizers.140 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.down_proj_weight_quantizers.141 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.down_proj_weight_quantizers.142 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.down_proj_weight_quantizers.143 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.50e-02 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.down_proj_weight_quantizers.144 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.64e-02 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.down_proj_weight_quantizers.145 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.down_proj_weight_quantizers.146 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.77e-01 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.down_proj_weight_quantizers.147 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.down_proj_weight_quantizers.148 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.down_proj_weight_quantizers.149 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.down_proj_weight_quantizers.150 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.down_proj_weight_quantizers.151 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.down_proj_weight_quantizers.152 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.down_proj_weight_quantizers.153 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.down_proj_weight_quantizers.154 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.down_proj_weight_quantizers.155 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.down_proj_weight_quantizers.156 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.down_proj_weight_quantizers.157 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.32e-01 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.down_proj_weight_quantizers.158 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.down_proj_weight_quantizers.159 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.down_proj_weight_quantizers.160 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.down_proj_weight_quantizers.161 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.31e-01 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.down_proj_weight_quantizers.162 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.down_proj_weight_quantizers.163 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.down_proj_weight_quantizers.164 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.54e-02 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.down_proj_weight_quantizers.165 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.down_proj_weight_quantizers.166 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.97e-01 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.down_proj_weight_quantizers.167 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.down_proj_weight_quantizers.168 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.down_proj_weight_quantizers.169 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.down_proj_weight_quantizers.170 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.down_proj_weight_quantizers.171 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.down_proj_weight_quantizers.172 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.16e-01 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.down_proj_weight_quantizers.173 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.44e-01 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.down_proj_weight_quantizers.174 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.66e-01 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.down_proj_weight_quantizers.175 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.down_proj_weight_quantizers.176 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.down_proj_weight_quantizers.177 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.down_proj_weight_quantizers.178 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.down_proj_weight_quantizers.179 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.51e-01 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.down_proj_weight_quantizers.180 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.down_proj_weight_quantizers.181 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.down_proj_weight_quantizers.182 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.down_proj_weight_quantizers.183 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.down_proj_weight_quantizers.184 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.down_proj_weight_quantizers.185 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.down_proj_weight_quantizers.186 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.down_proj_weight_quantizers.187 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.64e-01 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.down_proj_weight_quantizers.188 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.down_proj_weight_quantizers.189 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.down_proj_weight_quantizers.190 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.down_proj_weight_quantizers.191 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.down_proj_weight_quantizers.192 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.down_proj_weight_quantizers.193 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.down_proj_weight_quantizers.194 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.down_proj_weight_quantizers.195 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.down_proj_weight_quantizers.196 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.down_proj_weight_quantizers.197 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.down_proj_weight_quantizers.198 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.47e-01 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.down_proj_weight_quantizers.199 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.down_proj_weight_quantizers.200 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.85e-01 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.down_proj_weight_quantizers.201 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.down_proj_weight_quantizers.202 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.down_proj_weight_quantizers.203 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.46e-01 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.down_proj_weight_quantizers.204 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.down_proj_weight_quantizers.205 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.down_proj_weight_quantizers.206 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.down_proj_weight_quantizers.207 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.down_proj_weight_quantizers.208 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.down_proj_weight_quantizers.209 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.down_proj_weight_quantizers.210 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.down_proj_weight_quantizers.211 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.59e-02 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.down_proj_weight_quantizers.212 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.down_proj_weight_quantizers.213 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.25e-01 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.down_proj_weight_quantizers.214 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.down_proj_weight_quantizers.215 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.down_proj_weight_quantizers.216 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.24e-01 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.down_proj_weight_quantizers.217 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.down_proj_weight_quantizers.218 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.49e-01 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.down_proj_weight_quantizers.219 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.down_proj_weight_quantizers.220 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.down_proj_weight_quantizers.221 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.down_proj_weight_quantizers.222 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.down_proj_weight_quantizers.223 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.down_proj_weight_quantizers.224 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.down_proj_weight_quantizers.225 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.down_proj_weight_quantizers.226 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.down_proj_weight_quantizers.227 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.down_proj_weight_quantizers.228 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.down_proj_weight_quantizers.229 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.down_proj_weight_quantizers.230 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.59e-02 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.down_proj_weight_quantizers.231 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.down_proj_weight_quantizers.232 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.down_proj_weight_quantizers.233 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.down_proj_weight_quantizers.234 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.39e-01 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.down_proj_weight_quantizers.235 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.down_proj_weight_quantizers.236 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.down_proj_weight_quantizers.237 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.29e-01 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.down_proj_weight_quantizers.238 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.15e-01 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.down_proj_weight_quantizers.239 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.22e-01 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.down_proj_weight_quantizers.240 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.32e-01 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.down_proj_weight_quantizers.241 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.59e-02 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.down_proj_weight_quantizers.242 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.down_proj_weight_quantizers.243 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.down_proj_weight_quantizers.244 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.down_proj_weight_quantizers.245 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.down_proj_weight_quantizers.246 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.down_proj_weight_quantizers.247 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.34e-01 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.down_proj_weight_quantizers.248 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.down_proj_weight_quantizers.249 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.41e-01 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.down_proj_weight_quantizers.250 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.down_proj_weight_quantizers.251 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.down_proj_weight_quantizers.252 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.down_proj_weight_quantizers.253 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.down_proj_weight_quantizers.254 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.down_proj_weight_quantizers.255 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.40.mlp.shared_experts.gate_proj.input_quantizer TensorQuantizer(disabled) +model.layers.40.mlp.shared_experts.gate_proj.output_quantizer TensorQuantizer(disabled) +model.layers.40.mlp.shared_experts.gate_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.40.mlp.shared_experts.up_proj.input_quantizer TensorQuantizer(disabled) +model.layers.40.mlp.shared_experts.up_proj.output_quantizer TensorQuantizer(disabled) +model.layers.40.mlp.shared_experts.up_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.40.mlp.shared_experts.down_proj.input_quantizer TensorQuantizer(disabled) +model.layers.40.mlp.shared_experts.down_proj.output_quantizer TensorQuantizer(disabled) +model.layers.40.mlp.shared_experts.down_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.41.self_attn.q_a_proj.input_quantizer TensorQuantizer(disabled) +model.layers.41.self_attn.q_a_proj.output_quantizer TensorQuantizer(disabled) +model.layers.41.self_attn.q_a_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.41.self_attn.q_b_proj.input_quantizer TensorQuantizer(disabled) +model.layers.41.self_attn.q_b_proj.output_quantizer TensorQuantizer(disabled) +model.layers.41.self_attn.q_b_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.41.self_attn.kv_a_proj_with_mqa.input_quantizer TensorQuantizer(disabled) +model.layers.41.self_attn.kv_a_proj_with_mqa.output_quantizer TensorQuantizer(disabled) +model.layers.41.self_attn.kv_a_proj_with_mqa.weight_quantizer TensorQuantizer(disabled) +model.layers.41.self_attn.kv_b_proj.input_quantizer TensorQuantizer(disabled) +model.layers.41.self_attn.kv_b_proj.output_quantizer TensorQuantizer(disabled) +model.layers.41.self_attn.kv_b_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.41.self_attn.o_proj.input_quantizer TensorQuantizer(disabled) +model.layers.41.self_attn.o_proj.output_quantizer TensorQuantizer(disabled) +model.layers.41.self_attn.o_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.41.self_attn.q_bmm_quantizer TensorQuantizer(disabled) +model.layers.41.self_attn.k_bmm_quantizer TensorQuantizer((4, 3) bit fake per-tensor amax=4.48e+02(const) calibrator=MaxCalibrator quant) +model.layers.41.self_attn.v_bmm_quantizer TensorQuantizer((4, 3) bit fake per-tensor amax=4.48e+02(const) calibrator=MaxCalibrator quant) +model.layers.41.self_attn.p_bmm_quantizer TensorQuantizer(disabled) +model.layers.41.mlp.experts.gate_up_proj_input_quantizer TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.84e+00 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.gate_up_proj_weight_quantizers.0 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.gate_up_proj_weight_quantizers.1 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.gate_up_proj_weight_quantizers.2 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.13e-01 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.gate_up_proj_weight_quantizers.3 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.gate_up_proj_weight_quantizers.4 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.76e-01 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.gate_up_proj_weight_quantizers.5 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.68e-01 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.gate_up_proj_weight_quantizers.6 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.gate_up_proj_weight_quantizers.7 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.gate_up_proj_weight_quantizers.8 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.88e-01 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.gate_up_proj_weight_quantizers.9 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.16e-01 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.gate_up_proj_weight_quantizers.10 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.71e-01 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.gate_up_proj_weight_quantizers.11 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.68e-01 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.gate_up_proj_weight_quantizers.12 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.gate_up_proj_weight_quantizers.13 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.27e-01 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.gate_up_proj_weight_quantizers.14 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.gate_up_proj_weight_quantizers.15 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.gate_up_proj_weight_quantizers.16 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.50e-01 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.gate_up_proj_weight_quantizers.17 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.gate_up_proj_weight_quantizers.18 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.gate_up_proj_weight_quantizers.19 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.gate_up_proj_weight_quantizers.20 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.76e-01 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.gate_up_proj_weight_quantizers.21 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.gate_up_proj_weight_quantizers.22 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.gate_up_proj_weight_quantizers.23 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.gate_up_proj_weight_quantizers.24 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.gate_up_proj_weight_quantizers.25 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.gate_up_proj_weight_quantizers.26 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.gate_up_proj_weight_quantizers.27 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.19e-01 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.gate_up_proj_weight_quantizers.28 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.18e-01 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.gate_up_proj_weight_quantizers.29 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.gate_up_proj_weight_quantizers.30 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.17e-01 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.gate_up_proj_weight_quantizers.31 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.gate_up_proj_weight_quantizers.32 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.gate_up_proj_weight_quantizers.33 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.49e-01 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.gate_up_proj_weight_quantizers.34 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.17e-01 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.gate_up_proj_weight_quantizers.35 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.03e-01 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.gate_up_proj_weight_quantizers.36 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.19e-01 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.gate_up_proj_weight_quantizers.37 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.gate_up_proj_weight_quantizers.38 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.gate_up_proj_weight_quantizers.39 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.06e-01 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.gate_up_proj_weight_quantizers.40 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.58e-01 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.gate_up_proj_weight_quantizers.41 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.40e-01 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.gate_up_proj_weight_quantizers.42 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.gate_up_proj_weight_quantizers.43 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.58e-01 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.gate_up_proj_weight_quantizers.44 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.gate_up_proj_weight_quantizers.45 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.gate_up_proj_weight_quantizers.46 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=4.80e-01 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.gate_up_proj_weight_quantizers.47 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.gate_up_proj_weight_quantizers.48 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.23e-01 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.gate_up_proj_weight_quantizers.49 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.45e-01 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.gate_up_proj_weight_quantizers.50 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.17e-01 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.gate_up_proj_weight_quantizers.51 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.62e-01 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.gate_up_proj_weight_quantizers.52 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.gate_up_proj_weight_quantizers.53 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.15e-01 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.gate_up_proj_weight_quantizers.54 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.gate_up_proj_weight_quantizers.55 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.78e-01 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.gate_up_proj_weight_quantizers.56 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.80e-01 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.gate_up_proj_weight_quantizers.57 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.gate_up_proj_weight_quantizers.58 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.gate_up_proj_weight_quantizers.59 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.27e-01 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.gate_up_proj_weight_quantizers.60 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.gate_up_proj_weight_quantizers.61 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.16e-01 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.gate_up_proj_weight_quantizers.62 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.39e-01 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.gate_up_proj_weight_quantizers.63 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.gate_up_proj_weight_quantizers.64 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.56e-01 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.gate_up_proj_weight_quantizers.65 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.41e-01 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.gate_up_proj_weight_quantizers.66 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.41e-01 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.gate_up_proj_weight_quantizers.67 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.gate_up_proj_weight_quantizers.68 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.gate_up_proj_weight_quantizers.69 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.gate_up_proj_weight_quantizers.70 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=4.53e-01 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.gate_up_proj_weight_quantizers.71 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.30e-01 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.gate_up_proj_weight_quantizers.72 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.gate_up_proj_weight_quantizers.73 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.gate_up_proj_weight_quantizers.74 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.63e-01 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.gate_up_proj_weight_quantizers.75 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.gate_up_proj_weight_quantizers.76 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.24e-01 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.gate_up_proj_weight_quantizers.77 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.19e-01 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.gate_up_proj_weight_quantizers.78 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.gate_up_proj_weight_quantizers.79 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.99e-01 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.gate_up_proj_weight_quantizers.80 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.32e-01 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.gate_up_proj_weight_quantizers.81 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.26e-01 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.gate_up_proj_weight_quantizers.82 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.68e-01 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.gate_up_proj_weight_quantizers.83 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.gate_up_proj_weight_quantizers.84 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.60e-01 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.gate_up_proj_weight_quantizers.85 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.gate_up_proj_weight_quantizers.86 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.gate_up_proj_weight_quantizers.87 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.gate_up_proj_weight_quantizers.88 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.13e-01 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.gate_up_proj_weight_quantizers.89 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.gate_up_proj_weight_quantizers.90 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.gate_up_proj_weight_quantizers.91 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=5.43e-01 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.gate_up_proj_weight_quantizers.92 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.gate_up_proj_weight_quantizers.93 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.gate_up_proj_weight_quantizers.94 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.08e-01 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.gate_up_proj_weight_quantizers.95 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.27e-01 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.gate_up_proj_weight_quantizers.96 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.gate_up_proj_weight_quantizers.97 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.gate_up_proj_weight_quantizers.98 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.03e-01 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.gate_up_proj_weight_quantizers.99 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.gate_up_proj_weight_quantizers.100 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.34e-01 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.gate_up_proj_weight_quantizers.101 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.72e-01 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.gate_up_proj_weight_quantizers.102 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.gate_up_proj_weight_quantizers.103 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.26e-01 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.gate_up_proj_weight_quantizers.104 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.gate_up_proj_weight_quantizers.105 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.70e-01 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.gate_up_proj_weight_quantizers.106 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.gate_up_proj_weight_quantizers.107 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.32e-01 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.gate_up_proj_weight_quantizers.108 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.69e-01 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.gate_up_proj_weight_quantizers.109 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.88e-01 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.gate_up_proj_weight_quantizers.110 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.20e-01 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.gate_up_proj_weight_quantizers.111 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.gate_up_proj_weight_quantizers.112 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.60e-01 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.gate_up_proj_weight_quantizers.113 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.gate_up_proj_weight_quantizers.114 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.gate_up_proj_weight_quantizers.115 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.gate_up_proj_weight_quantizers.116 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.26e-01 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.gate_up_proj_weight_quantizers.117 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.gate_up_proj_weight_quantizers.118 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.88e-01 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.gate_up_proj_weight_quantizers.119 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.gate_up_proj_weight_quantizers.120 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.gate_up_proj_weight_quantizers.121 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.58e-01 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.gate_up_proj_weight_quantizers.122 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.86e-01 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.gate_up_proj_weight_quantizers.123 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.gate_up_proj_weight_quantizers.124 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.gate_up_proj_weight_quantizers.125 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.56e-01 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.gate_up_proj_weight_quantizers.126 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.14e-01 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.gate_up_proj_weight_quantizers.127 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.19e-01 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.gate_up_proj_weight_quantizers.128 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.gate_up_proj_weight_quantizers.129 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.30e-01 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.gate_up_proj_weight_quantizers.130 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.gate_up_proj_weight_quantizers.131 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.52e-01 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.gate_up_proj_weight_quantizers.132 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.79e-01 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.gate_up_proj_weight_quantizers.133 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.39e-01 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.gate_up_proj_weight_quantizers.134 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.37e-01 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.gate_up_proj_weight_quantizers.135 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.27e-01 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.gate_up_proj_weight_quantizers.136 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.gate_up_proj_weight_quantizers.137 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.gate_up_proj_weight_quantizers.138 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.01e-01 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.gate_up_proj_weight_quantizers.139 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.92e-01 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.gate_up_proj_weight_quantizers.140 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.18e-01 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.gate_up_proj_weight_quantizers.141 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.14e-01 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.gate_up_proj_weight_quantizers.142 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.gate_up_proj_weight_quantizers.143 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.gate_up_proj_weight_quantizers.144 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.16e-01 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.gate_up_proj_weight_quantizers.145 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.22e-01 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.gate_up_proj_weight_quantizers.146 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.gate_up_proj_weight_quantizers.147 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.29e-01 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.gate_up_proj_weight_quantizers.148 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.gate_up_proj_weight_quantizers.149 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.71e-01 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.gate_up_proj_weight_quantizers.150 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.gate_up_proj_weight_quantizers.151 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.89e-01 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.gate_up_proj_weight_quantizers.152 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.53e-01 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.gate_up_proj_weight_quantizers.153 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.27e-01 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.gate_up_proj_weight_quantizers.154 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.84e-01 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.gate_up_proj_weight_quantizers.155 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.58e-01 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.gate_up_proj_weight_quantizers.156 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.20e-01 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.gate_up_proj_weight_quantizers.157 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.gate_up_proj_weight_quantizers.158 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.29e-01 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.gate_up_proj_weight_quantizers.159 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.gate_up_proj_weight_quantizers.160 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.gate_up_proj_weight_quantizers.161 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.25e-01 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.gate_up_proj_weight_quantizers.162 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.41e-01 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.gate_up_proj_weight_quantizers.163 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.gate_up_proj_weight_quantizers.164 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.gate_up_proj_weight_quantizers.165 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.gate_up_proj_weight_quantizers.166 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.17e-01 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.gate_up_proj_weight_quantizers.167 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.gate_up_proj_weight_quantizers.168 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.22e-01 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.gate_up_proj_weight_quantizers.169 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.gate_up_proj_weight_quantizers.170 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.gate_up_proj_weight_quantizers.171 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.gate_up_proj_weight_quantizers.172 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.gate_up_proj_weight_quantizers.173 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.25e-01 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.gate_up_proj_weight_quantizers.174 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.gate_up_proj_weight_quantizers.175 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.15e-01 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.gate_up_proj_weight_quantizers.176 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.gate_up_proj_weight_quantizers.177 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.15e-01 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.gate_up_proj_weight_quantizers.178 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.11e-01 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.gate_up_proj_weight_quantizers.179 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.gate_up_proj_weight_quantizers.180 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.29e-01 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.gate_up_proj_weight_quantizers.181 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.gate_up_proj_weight_quantizers.182 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.66e-01 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.gate_up_proj_weight_quantizers.183 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.gate_up_proj_weight_quantizers.184 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.gate_up_proj_weight_quantizers.185 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.20e-01 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.gate_up_proj_weight_quantizers.186 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.gate_up_proj_weight_quantizers.187 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.71e-01 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.gate_up_proj_weight_quantizers.188 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.gate_up_proj_weight_quantizers.189 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.gate_up_proj_weight_quantizers.190 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.15e-01 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.gate_up_proj_weight_quantizers.191 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.gate_up_proj_weight_quantizers.192 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.gate_up_proj_weight_quantizers.193 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.gate_up_proj_weight_quantizers.194 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.40e-01 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.gate_up_proj_weight_quantizers.195 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.gate_up_proj_weight_quantizers.196 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.40e-01 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.gate_up_proj_weight_quantizers.197 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.26e-01 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.gate_up_proj_weight_quantizers.198 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.gate_up_proj_weight_quantizers.199 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.18e-01 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.gate_up_proj_weight_quantizers.200 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.gate_up_proj_weight_quantizers.201 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.gate_up_proj_weight_quantizers.202 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.46e-01 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.gate_up_proj_weight_quantizers.203 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.55e-01 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.gate_up_proj_weight_quantizers.204 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.gate_up_proj_weight_quantizers.205 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.22e-01 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.gate_up_proj_weight_quantizers.206 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.gate_up_proj_weight_quantizers.207 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.gate_up_proj_weight_quantizers.208 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.22e-01 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.gate_up_proj_weight_quantizers.209 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.41e-01 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.gate_up_proj_weight_quantizers.210 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.gate_up_proj_weight_quantizers.211 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.gate_up_proj_weight_quantizers.212 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.gate_up_proj_weight_quantizers.213 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.17e-01 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.gate_up_proj_weight_quantizers.214 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.21e-01 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.gate_up_proj_weight_quantizers.215 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.gate_up_proj_weight_quantizers.216 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.48e-01 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.gate_up_proj_weight_quantizers.217 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.23e-01 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.gate_up_proj_weight_quantizers.218 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.22e-01 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.gate_up_proj_weight_quantizers.219 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.94e-01 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.gate_up_proj_weight_quantizers.220 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.gate_up_proj_weight_quantizers.221 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.gate_up_proj_weight_quantizers.222 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.gate_up_proj_weight_quantizers.223 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.14e-01 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.gate_up_proj_weight_quantizers.224 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.12e-01 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.gate_up_proj_weight_quantizers.225 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.37e-01 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.gate_up_proj_weight_quantizers.226 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.93e-01 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.gate_up_proj_weight_quantizers.227 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.29e-01 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.gate_up_proj_weight_quantizers.228 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=4.14e-01 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.gate_up_proj_weight_quantizers.229 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.23e-01 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.gate_up_proj_weight_quantizers.230 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.37e-01 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.gate_up_proj_weight_quantizers.231 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.gate_up_proj_weight_quantizers.232 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.36e-01 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.gate_up_proj_weight_quantizers.233 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.gate_up_proj_weight_quantizers.234 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.gate_up_proj_weight_quantizers.235 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.gate_up_proj_weight_quantizers.236 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.gate_up_proj_weight_quantizers.237 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.gate_up_proj_weight_quantizers.238 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.70e-01 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.gate_up_proj_weight_quantizers.239 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.gate_up_proj_weight_quantizers.240 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.71e-01 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.gate_up_proj_weight_quantizers.241 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.gate_up_proj_weight_quantizers.242 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.75e-01 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.gate_up_proj_weight_quantizers.243 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.gate_up_proj_weight_quantizers.244 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.39e-01 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.gate_up_proj_weight_quantizers.245 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.gate_up_proj_weight_quantizers.246 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.gate_up_proj_weight_quantizers.247 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.52e-01 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.gate_up_proj_weight_quantizers.248 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.36e-01 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.gate_up_proj_weight_quantizers.249 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.gate_up_proj_weight_quantizers.250 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.gate_up_proj_weight_quantizers.251 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.gate_up_proj_weight_quantizers.252 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.gate_up_proj_weight_quantizers.253 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.gate_up_proj_weight_quantizers.254 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.gate_up_proj_weight_quantizers.255 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.18e-01 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.down_proj_input_quantizer TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.38e+01 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.down_proj_weight_quantizers.0 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.down_proj_weight_quantizers.1 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.down_proj_weight_quantizers.2 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.down_proj_weight_quantizers.3 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.down_proj_weight_quantizers.4 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.down_proj_weight_quantizers.5 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.down_proj_weight_quantizers.6 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.down_proj_weight_quantizers.7 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.down_proj_weight_quantizers.8 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.down_proj_weight_quantizers.9 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.down_proj_weight_quantizers.10 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.down_proj_weight_quantizers.11 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.down_proj_weight_quantizers.12 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.down_proj_weight_quantizers.13 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.down_proj_weight_quantizers.14 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.down_proj_weight_quantizers.15 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.down_proj_weight_quantizers.16 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.down_proj_weight_quantizers.17 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.down_proj_weight_quantizers.18 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.54e-02 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.down_proj_weight_quantizers.19 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.down_proj_weight_quantizers.20 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.54e-02 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.down_proj_weight_quantizers.21 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.down_proj_weight_quantizers.22 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.down_proj_weight_quantizers.23 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.down_proj_weight_quantizers.24 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.down_proj_weight_quantizers.25 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.down_proj_weight_quantizers.26 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.down_proj_weight_quantizers.27 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.down_proj_weight_quantizers.28 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.down_proj_weight_quantizers.29 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.down_proj_weight_quantizers.30 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.53e-01 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.down_proj_weight_quantizers.31 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.down_proj_weight_quantizers.32 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.down_proj_weight_quantizers.33 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.30e-02 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.down_proj_weight_quantizers.34 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.down_proj_weight_quantizers.35 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=4.94e-01 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.down_proj_weight_quantizers.36 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.down_proj_weight_quantizers.37 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.down_proj_weight_quantizers.38 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.23e-01 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.down_proj_weight_quantizers.39 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.30e-01 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.down_proj_weight_quantizers.40 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.down_proj_weight_quantizers.41 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.down_proj_weight_quantizers.42 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.down_proj_weight_quantizers.43 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.35e-01 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.down_proj_weight_quantizers.44 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.down_proj_weight_quantizers.45 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.down_proj_weight_quantizers.46 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.03e-01 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.down_proj_weight_quantizers.47 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.down_proj_weight_quantizers.48 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.down_proj_weight_quantizers.49 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.down_proj_weight_quantizers.50 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.40e-02 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.down_proj_weight_quantizers.51 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.19e-01 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.down_proj_weight_quantizers.52 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.down_proj_weight_quantizers.53 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.down_proj_weight_quantizers.54 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.down_proj_weight_quantizers.55 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.down_proj_weight_quantizers.56 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.50e-02 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.down_proj_weight_quantizers.57 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.down_proj_weight_quantizers.58 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.59e-02 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.down_proj_weight_quantizers.59 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.15e-01 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.down_proj_weight_quantizers.60 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.down_proj_weight_quantizers.61 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.19e-01 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.down_proj_weight_quantizers.62 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.14e-01 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.down_proj_weight_quantizers.63 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.down_proj_weight_quantizers.64 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.down_proj_weight_quantizers.65 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.down_proj_weight_quantizers.66 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.down_proj_weight_quantizers.67 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.down_proj_weight_quantizers.68 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.59e-02 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.down_proj_weight_quantizers.69 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.down_proj_weight_quantizers.70 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.87e-01 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.down_proj_weight_quantizers.71 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.down_proj_weight_quantizers.72 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.down_proj_weight_quantizers.73 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.down_proj_weight_quantizers.74 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.50e-01 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.down_proj_weight_quantizers.75 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.down_proj_weight_quantizers.76 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.down_proj_weight_quantizers.77 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.down_proj_weight_quantizers.78 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.down_proj_weight_quantizers.79 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.down_proj_weight_quantizers.80 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.down_proj_weight_quantizers.81 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.down_proj_weight_quantizers.82 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.down_proj_weight_quantizers.83 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.down_proj_weight_quantizers.84 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.31e-01 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.down_proj_weight_quantizers.85 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.down_proj_weight_quantizers.86 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.down_proj_weight_quantizers.87 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.down_proj_weight_quantizers.88 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.down_proj_weight_quantizers.89 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.down_proj_weight_quantizers.90 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.54e-02 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.down_proj_weight_quantizers.91 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.95e-01 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.down_proj_weight_quantizers.92 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.down_proj_weight_quantizers.93 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.down_proj_weight_quantizers.94 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.88e-01 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.down_proj_weight_quantizers.95 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.down_proj_weight_quantizers.96 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.down_proj_weight_quantizers.97 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.75e-01 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.down_proj_weight_quantizers.98 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.37e-01 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.down_proj_weight_quantizers.99 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.down_proj_weight_quantizers.100 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.down_proj_weight_quantizers.101 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.down_proj_weight_quantizers.102 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.down_proj_weight_quantizers.103 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.88e-01 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.down_proj_weight_quantizers.104 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.down_proj_weight_quantizers.105 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.down_proj_weight_quantizers.106 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.down_proj_weight_quantizers.107 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.down_proj_weight_quantizers.108 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.down_proj_weight_quantizers.109 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.32e-01 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.down_proj_weight_quantizers.110 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.down_proj_weight_quantizers.111 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.down_proj_weight_quantizers.112 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.down_proj_weight_quantizers.113 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.down_proj_weight_quantizers.114 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.down_proj_weight_quantizers.115 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.down_proj_weight_quantizers.116 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.down_proj_weight_quantizers.117 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.down_proj_weight_quantizers.118 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.down_proj_weight_quantizers.119 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.down_proj_weight_quantizers.120 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.down_proj_weight_quantizers.121 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.28e-01 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.down_proj_weight_quantizers.122 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.down_proj_weight_quantizers.123 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.down_proj_weight_quantizers.124 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.down_proj_weight_quantizers.125 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.63e-01 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.down_proj_weight_quantizers.126 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.59e-02 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.down_proj_weight_quantizers.127 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.31e-01 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.down_proj_weight_quantizers.128 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.54e-02 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.down_proj_weight_quantizers.129 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.down_proj_weight_quantizers.130 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.down_proj_weight_quantizers.131 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.43e-01 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.down_proj_weight_quantizers.132 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.down_proj_weight_quantizers.133 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.down_proj_weight_quantizers.134 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.down_proj_weight_quantizers.135 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.down_proj_weight_quantizers.136 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.down_proj_weight_quantizers.137 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.down_proj_weight_quantizers.138 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.down_proj_weight_quantizers.139 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.down_proj_weight_quantizers.140 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.50e-02 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.down_proj_weight_quantizers.141 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.down_proj_weight_quantizers.142 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.down_proj_weight_quantizers.143 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.down_proj_weight_quantizers.144 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.down_proj_weight_quantizers.145 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.down_proj_weight_quantizers.146 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.down_proj_weight_quantizers.147 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.down_proj_weight_quantizers.148 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.down_proj_weight_quantizers.149 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.59e-02 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.down_proj_weight_quantizers.150 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.down_proj_weight_quantizers.151 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.down_proj_weight_quantizers.152 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.down_proj_weight_quantizers.153 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.down_proj_weight_quantizers.154 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.22e-01 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.down_proj_weight_quantizers.155 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.32e-01 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.down_proj_weight_quantizers.156 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.down_proj_weight_quantizers.157 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.down_proj_weight_quantizers.158 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.61e-01 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.down_proj_weight_quantizers.159 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.50e-02 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.down_proj_weight_quantizers.160 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.down_proj_weight_quantizers.161 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.down_proj_weight_quantizers.162 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.down_proj_weight_quantizers.163 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.down_proj_weight_quantizers.164 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.down_proj_weight_quantizers.165 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.54e-02 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.down_proj_weight_quantizers.166 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.down_proj_weight_quantizers.167 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.down_proj_weight_quantizers.168 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.95e-01 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.down_proj_weight_quantizers.169 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.down_proj_weight_quantizers.170 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.down_proj_weight_quantizers.171 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.down_proj_weight_quantizers.172 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.64e-02 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.down_proj_weight_quantizers.173 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.down_proj_weight_quantizers.174 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.down_proj_weight_quantizers.175 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.50e-02 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.down_proj_weight_quantizers.176 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.down_proj_weight_quantizers.177 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.24e-01 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.down_proj_weight_quantizers.178 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.45e-02 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.down_proj_weight_quantizers.179 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.down_proj_weight_quantizers.180 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.down_proj_weight_quantizers.181 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.down_proj_weight_quantizers.182 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.45e-02 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.down_proj_weight_quantizers.183 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.down_proj_weight_quantizers.184 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.40e-02 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.down_proj_weight_quantizers.185 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.79e-01 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.down_proj_weight_quantizers.186 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.64e-02 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.down_proj_weight_quantizers.187 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.55e-01 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.down_proj_weight_quantizers.188 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.down_proj_weight_quantizers.189 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.50e-02 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.down_proj_weight_quantizers.190 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.31e-01 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.down_proj_weight_quantizers.191 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.down_proj_weight_quantizers.192 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.down_proj_weight_quantizers.193 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.down_proj_weight_quantizers.194 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.down_proj_weight_quantizers.195 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.down_proj_weight_quantizers.196 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.46e-01 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.down_proj_weight_quantizers.197 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.down_proj_weight_quantizers.198 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.down_proj_weight_quantizers.199 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.down_proj_weight_quantizers.200 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.down_proj_weight_quantizers.201 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.down_proj_weight_quantizers.202 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.down_proj_weight_quantizers.203 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.26e-01 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.down_proj_weight_quantizers.204 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.45e-02 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.down_proj_weight_quantizers.205 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.64e-02 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.down_proj_weight_quantizers.206 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.down_proj_weight_quantizers.207 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.down_proj_weight_quantizers.208 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.down_proj_weight_quantizers.209 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.down_proj_weight_quantizers.210 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.down_proj_weight_quantizers.211 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.down_proj_weight_quantizers.212 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.down_proj_weight_quantizers.213 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.down_proj_weight_quantizers.214 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.down_proj_weight_quantizers.215 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.down_proj_weight_quantizers.216 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.14e-01 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.down_proj_weight_quantizers.217 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.down_proj_weight_quantizers.218 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.25e-02 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.down_proj_weight_quantizers.219 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.52e-01 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.down_proj_weight_quantizers.220 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.down_proj_weight_quantizers.221 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.down_proj_weight_quantizers.222 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.down_proj_weight_quantizers.223 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.down_proj_weight_quantizers.224 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.down_proj_weight_quantizers.225 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.19e-01 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.down_proj_weight_quantizers.226 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.27e-01 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.down_proj_weight_quantizers.227 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.down_proj_weight_quantizers.228 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.41e-01 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.down_proj_weight_quantizers.229 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.down_proj_weight_quantizers.230 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.down_proj_weight_quantizers.231 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.down_proj_weight_quantizers.232 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.down_proj_weight_quantizers.233 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.down_proj_weight_quantizers.234 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.45e-02 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.down_proj_weight_quantizers.235 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.down_proj_weight_quantizers.236 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.down_proj_weight_quantizers.237 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.down_proj_weight_quantizers.238 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.down_proj_weight_quantizers.239 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.down_proj_weight_quantizers.240 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.down_proj_weight_quantizers.241 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.down_proj_weight_quantizers.242 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.31e-01 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.down_proj_weight_quantizers.243 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.down_proj_weight_quantizers.244 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.down_proj_weight_quantizers.245 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.down_proj_weight_quantizers.246 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.down_proj_weight_quantizers.247 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.53e-01 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.down_proj_weight_quantizers.248 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.60e-01 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.down_proj_weight_quantizers.249 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.down_proj_weight_quantizers.250 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.down_proj_weight_quantizers.251 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.down_proj_weight_quantizers.252 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.down_proj_weight_quantizers.253 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.down_proj_weight_quantizers.254 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.down_proj_weight_quantizers.255 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.83e-01 calibrator=MaxCalibrator quant) +model.layers.41.mlp.shared_experts.gate_proj.input_quantizer TensorQuantizer(disabled) +model.layers.41.mlp.shared_experts.gate_proj.output_quantizer TensorQuantizer(disabled) +model.layers.41.mlp.shared_experts.gate_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.41.mlp.shared_experts.up_proj.input_quantizer TensorQuantizer(disabled) +model.layers.41.mlp.shared_experts.up_proj.output_quantizer TensorQuantizer(disabled) +model.layers.41.mlp.shared_experts.up_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.41.mlp.shared_experts.down_proj.input_quantizer TensorQuantizer(disabled) +model.layers.41.mlp.shared_experts.down_proj.output_quantizer TensorQuantizer(disabled) +model.layers.41.mlp.shared_experts.down_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.42.self_attn.q_a_proj.input_quantizer TensorQuantizer(disabled) +model.layers.42.self_attn.q_a_proj.output_quantizer TensorQuantizer(disabled) +model.layers.42.self_attn.q_a_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.42.self_attn.q_b_proj.input_quantizer TensorQuantizer(disabled) +model.layers.42.self_attn.q_b_proj.output_quantizer TensorQuantizer(disabled) +model.layers.42.self_attn.q_b_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.42.self_attn.kv_a_proj_with_mqa.input_quantizer TensorQuantizer(disabled) +model.layers.42.self_attn.kv_a_proj_with_mqa.output_quantizer TensorQuantizer(disabled) +model.layers.42.self_attn.kv_a_proj_with_mqa.weight_quantizer TensorQuantizer(disabled) +model.layers.42.self_attn.kv_b_proj.input_quantizer TensorQuantizer(disabled) +model.layers.42.self_attn.kv_b_proj.output_quantizer TensorQuantizer(disabled) +model.layers.42.self_attn.kv_b_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.42.self_attn.o_proj.input_quantizer TensorQuantizer(disabled) +model.layers.42.self_attn.o_proj.output_quantizer TensorQuantizer(disabled) +model.layers.42.self_attn.o_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.42.self_attn.indexer.wq_b.input_quantizer TensorQuantizer(disabled) +model.layers.42.self_attn.indexer.wq_b.output_quantizer TensorQuantizer(disabled) +model.layers.42.self_attn.indexer.wq_b.weight_quantizer TensorQuantizer(disabled) +model.layers.42.self_attn.indexer.wk.input_quantizer TensorQuantizer(disabled) +model.layers.42.self_attn.indexer.wk.output_quantizer TensorQuantizer(disabled) +model.layers.42.self_attn.indexer.wk.weight_quantizer TensorQuantizer(disabled) +model.layers.42.self_attn.indexer.k_norm.input_quantizer TensorQuantizer(disabled) +model.layers.42.self_attn.indexer.k_norm.output_quantizer TensorQuantizer(disabled) +model.layers.42.self_attn.indexer.weights_proj.input_quantizer TensorQuantizer(disabled) +model.layers.42.self_attn.indexer.weights_proj.output_quantizer TensorQuantizer(disabled) +model.layers.42.self_attn.indexer.weights_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.42.self_attn.q_bmm_quantizer TensorQuantizer(disabled) +model.layers.42.self_attn.k_bmm_quantizer TensorQuantizer((4, 3) bit fake per-tensor amax=4.48e+02(const) calibrator=MaxCalibrator quant) +model.layers.42.self_attn.v_bmm_quantizer TensorQuantizer((4, 3) bit fake per-tensor amax=4.48e+02(const) calibrator=MaxCalibrator quant) +model.layers.42.self_attn.p_bmm_quantizer TensorQuantizer(disabled) +model.layers.42.mlp.experts.gate_up_proj_input_quantizer TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.05e+00 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.gate_up_proj_weight_quantizers.0 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.15e-01 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.gate_up_proj_weight_quantizers.1 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.gate_up_proj_weight_quantizers.2 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.70e-01 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.gate_up_proj_weight_quantizers.3 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.39e-01 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.gate_up_proj_weight_quantizers.4 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.gate_up_proj_weight_quantizers.5 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.gate_up_proj_weight_quantizers.6 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.gate_up_proj_weight_quantizers.7 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.75e-01 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.gate_up_proj_weight_quantizers.8 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.81e-01 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.gate_up_proj_weight_quantizers.9 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.18e-01 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.gate_up_proj_weight_quantizers.10 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.22e-01 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.gate_up_proj_weight_quantizers.11 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.71e-01 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.gate_up_proj_weight_quantizers.12 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.gate_up_proj_weight_quantizers.13 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.gate_up_proj_weight_quantizers.14 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.gate_up_proj_weight_quantizers.15 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.gate_up_proj_weight_quantizers.16 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.17e-01 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.gate_up_proj_weight_quantizers.17 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.15e-01 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.gate_up_proj_weight_quantizers.18 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.73e-01 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.gate_up_proj_weight_quantizers.19 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.18e-01 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.gate_up_proj_weight_quantizers.20 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.16e-01 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.gate_up_proj_weight_quantizers.21 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=4.28e-01 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.gate_up_proj_weight_quantizers.22 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.25e-01 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.gate_up_proj_weight_quantizers.23 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.gate_up_proj_weight_quantizers.24 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.gate_up_proj_weight_quantizers.25 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.gate_up_proj_weight_quantizers.26 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.56e-01 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.gate_up_proj_weight_quantizers.27 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.gate_up_proj_weight_quantizers.28 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.gate_up_proj_weight_quantizers.29 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.gate_up_proj_weight_quantizers.30 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.gate_up_proj_weight_quantizers.31 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.gate_up_proj_weight_quantizers.32 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.23e-01 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.gate_up_proj_weight_quantizers.33 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.gate_up_proj_weight_quantizers.34 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.gate_up_proj_weight_quantizers.35 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.50e-01 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.gate_up_proj_weight_quantizers.36 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.65e-01 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.gate_up_proj_weight_quantizers.37 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.73e-01 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.gate_up_proj_weight_quantizers.38 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.26e-01 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.gate_up_proj_weight_quantizers.39 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.gate_up_proj_weight_quantizers.40 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.gate_up_proj_weight_quantizers.41 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.gate_up_proj_weight_quantizers.42 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.gate_up_proj_weight_quantizers.43 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.gate_up_proj_weight_quantizers.44 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.52e-01 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.gate_up_proj_weight_quantizers.45 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.gate_up_proj_weight_quantizers.46 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.gate_up_proj_weight_quantizers.47 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.42e-01 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.gate_up_proj_weight_quantizers.48 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.gate_up_proj_weight_quantizers.49 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.33e-01 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.gate_up_proj_weight_quantizers.50 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.gate_up_proj_weight_quantizers.51 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.91e-01 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.gate_up_proj_weight_quantizers.52 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.gate_up_proj_weight_quantizers.53 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.29e-01 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.gate_up_proj_weight_quantizers.54 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.gate_up_proj_weight_quantizers.55 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.gate_up_proj_weight_quantizers.56 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.80e-01 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.gate_up_proj_weight_quantizers.57 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.54e-02 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.gate_up_proj_weight_quantizers.58 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.gate_up_proj_weight_quantizers.59 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.gate_up_proj_weight_quantizers.60 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.gate_up_proj_weight_quantizers.61 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.17e-01 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.gate_up_proj_weight_quantizers.62 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.gate_up_proj_weight_quantizers.63 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.gate_up_proj_weight_quantizers.64 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.gate_up_proj_weight_quantizers.65 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.gate_up_proj_weight_quantizers.66 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.gate_up_proj_weight_quantizers.67 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.31e-01 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.gate_up_proj_weight_quantizers.68 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.35e-01 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.gate_up_proj_weight_quantizers.69 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.56e-01 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.gate_up_proj_weight_quantizers.70 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.gate_up_proj_weight_quantizers.71 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.81e-01 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.gate_up_proj_weight_quantizers.72 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.gate_up_proj_weight_quantizers.73 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.gate_up_proj_weight_quantizers.74 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.35e-01 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.gate_up_proj_weight_quantizers.75 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.gate_up_proj_weight_quantizers.76 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.gate_up_proj_weight_quantizers.77 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.45e-01 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.gate_up_proj_weight_quantizers.78 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.gate_up_proj_weight_quantizers.79 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.gate_up_proj_weight_quantizers.80 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=4.00e-01 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.gate_up_proj_weight_quantizers.81 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.gate_up_proj_weight_quantizers.82 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.45e-01 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.gate_up_proj_weight_quantizers.83 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.gate_up_proj_weight_quantizers.84 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.gate_up_proj_weight_quantizers.85 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.95e-01 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.gate_up_proj_weight_quantizers.86 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.73e-01 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.gate_up_proj_weight_quantizers.87 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.25e-01 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.gate_up_proj_weight_quantizers.88 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.14e-01 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.gate_up_proj_weight_quantizers.89 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.gate_up_proj_weight_quantizers.90 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.29e-01 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.gate_up_proj_weight_quantizers.91 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.81e-01 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.gate_up_proj_weight_quantizers.92 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.37e-01 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.gate_up_proj_weight_quantizers.93 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.61e-01 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.gate_up_proj_weight_quantizers.94 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.gate_up_proj_weight_quantizers.95 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.gate_up_proj_weight_quantizers.96 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.63e-01 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.gate_up_proj_weight_quantizers.97 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.60e-01 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.gate_up_proj_weight_quantizers.98 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.gate_up_proj_weight_quantizers.99 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.36e-01 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.gate_up_proj_weight_quantizers.100 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.31e-01 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.gate_up_proj_weight_quantizers.101 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.22e-01 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.gate_up_proj_weight_quantizers.102 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.34e-01 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.gate_up_proj_weight_quantizers.103 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.03e-01 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.gate_up_proj_weight_quantizers.104 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.72e-01 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.gate_up_proj_weight_quantizers.105 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.gate_up_proj_weight_quantizers.106 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.gate_up_proj_weight_quantizers.107 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.01e-01 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.gate_up_proj_weight_quantizers.108 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.gate_up_proj_weight_quantizers.109 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.gate_up_proj_weight_quantizers.110 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.23e-01 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.gate_up_proj_weight_quantizers.111 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.gate_up_proj_weight_quantizers.112 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.84e-01 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.gate_up_proj_weight_quantizers.113 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.gate_up_proj_weight_quantizers.114 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.gate_up_proj_weight_quantizers.115 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.gate_up_proj_weight_quantizers.116 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.71e-01 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.gate_up_proj_weight_quantizers.117 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.33e-01 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.gate_up_proj_weight_quantizers.118 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.25e-01 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.gate_up_proj_weight_quantizers.119 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.16e-01 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.gate_up_proj_weight_quantizers.120 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.gate_up_proj_weight_quantizers.121 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.gate_up_proj_weight_quantizers.122 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.14e-01 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.gate_up_proj_weight_quantizers.123 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.88e-01 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.gate_up_proj_weight_quantizers.124 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.27e-01 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.gate_up_proj_weight_quantizers.125 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.25e-01 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.gate_up_proj_weight_quantizers.126 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.45e-01 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.gate_up_proj_weight_quantizers.127 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.22e-01 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.gate_up_proj_weight_quantizers.128 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.26e-01 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.gate_up_proj_weight_quantizers.129 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.gate_up_proj_weight_quantizers.130 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.60e-01 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.gate_up_proj_weight_quantizers.131 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.14e-01 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.gate_up_proj_weight_quantizers.132 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.gate_up_proj_weight_quantizers.133 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.18e-01 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.gate_up_proj_weight_quantizers.134 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.gate_up_proj_weight_quantizers.135 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.gate_up_proj_weight_quantizers.136 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.gate_up_proj_weight_quantizers.137 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.56e-01 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.gate_up_proj_weight_quantizers.138 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.20e-01 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.gate_up_proj_weight_quantizers.139 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.gate_up_proj_weight_quantizers.140 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.gate_up_proj_weight_quantizers.141 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.42e-01 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.gate_up_proj_weight_quantizers.142 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.gate_up_proj_weight_quantizers.143 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.80e-01 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.gate_up_proj_weight_quantizers.144 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.08e-01 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.gate_up_proj_weight_quantizers.145 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.34e-01 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.gate_up_proj_weight_quantizers.146 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.gate_up_proj_weight_quantizers.147 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.gate_up_proj_weight_quantizers.148 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.45e-01 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.gate_up_proj_weight_quantizers.149 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.gate_up_proj_weight_quantizers.150 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.09e-01 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.gate_up_proj_weight_quantizers.151 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.gate_up_proj_weight_quantizers.152 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.gate_up_proj_weight_quantizers.153 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.08e-01 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.gate_up_proj_weight_quantizers.154 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.gate_up_proj_weight_quantizers.155 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.gate_up_proj_weight_quantizers.156 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.31e-01 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.gate_up_proj_weight_quantizers.157 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.20e-01 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.gate_up_proj_weight_quantizers.158 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.gate_up_proj_weight_quantizers.159 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.gate_up_proj_weight_quantizers.160 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.gate_up_proj_weight_quantizers.161 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.94e-01 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.gate_up_proj_weight_quantizers.162 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.65e-01 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.gate_up_proj_weight_quantizers.163 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.26e-01 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.gate_up_proj_weight_quantizers.164 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.47e-01 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.gate_up_proj_weight_quantizers.165 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.gate_up_proj_weight_quantizers.166 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.gate_up_proj_weight_quantizers.167 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.24e-01 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.gate_up_proj_weight_quantizers.168 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.gate_up_proj_weight_quantizers.169 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.gate_up_proj_weight_quantizers.170 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.17e-01 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.gate_up_proj_weight_quantizers.171 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.gate_up_proj_weight_quantizers.172 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.22e-01 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.gate_up_proj_weight_quantizers.173 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.gate_up_proj_weight_quantizers.174 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.gate_up_proj_weight_quantizers.175 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.gate_up_proj_weight_quantizers.176 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.gate_up_proj_weight_quantizers.177 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.83e-01 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.gate_up_proj_weight_quantizers.178 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.gate_up_proj_weight_quantizers.179 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.gate_up_proj_weight_quantizers.180 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.88e-01 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.gate_up_proj_weight_quantizers.181 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.gate_up_proj_weight_quantizers.182 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.43e-01 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.gate_up_proj_weight_quantizers.183 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.16e-01 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.gate_up_proj_weight_quantizers.184 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.gate_up_proj_weight_quantizers.185 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.gate_up_proj_weight_quantizers.186 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.gate_up_proj_weight_quantizers.187 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.gate_up_proj_weight_quantizers.188 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.55e-01 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.gate_up_proj_weight_quantizers.189 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.19e-01 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.gate_up_proj_weight_quantizers.190 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.gate_up_proj_weight_quantizers.191 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.gate_up_proj_weight_quantizers.192 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.gate_up_proj_weight_quantizers.193 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.gate_up_proj_weight_quantizers.194 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.20e-01 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.gate_up_proj_weight_quantizers.195 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.gate_up_proj_weight_quantizers.196 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.gate_up_proj_weight_quantizers.197 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.gate_up_proj_weight_quantizers.198 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.41e-01 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.gate_up_proj_weight_quantizers.199 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.gate_up_proj_weight_quantizers.200 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.66e-01 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.gate_up_proj_weight_quantizers.201 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.gate_up_proj_weight_quantizers.202 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.gate_up_proj_weight_quantizers.203 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.gate_up_proj_weight_quantizers.204 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.gate_up_proj_weight_quantizers.205 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.95e-01 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.gate_up_proj_weight_quantizers.206 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.gate_up_proj_weight_quantizers.207 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.41e-01 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.gate_up_proj_weight_quantizers.208 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=4.02e-01 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.gate_up_proj_weight_quantizers.209 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.gate_up_proj_weight_quantizers.210 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.gate_up_proj_weight_quantizers.211 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.18e-01 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.gate_up_proj_weight_quantizers.212 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.02e-01 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.gate_up_proj_weight_quantizers.213 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.23e-01 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.gate_up_proj_weight_quantizers.214 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.gate_up_proj_weight_quantizers.215 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.16e-01 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.gate_up_proj_weight_quantizers.216 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.45e-01 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.gate_up_proj_weight_quantizers.217 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.gate_up_proj_weight_quantizers.218 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.79e-01 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.gate_up_proj_weight_quantizers.219 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.13e-01 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.gate_up_proj_weight_quantizers.220 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.gate_up_proj_weight_quantizers.221 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=4.10e-01 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.gate_up_proj_weight_quantizers.222 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.46e-01 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.gate_up_proj_weight_quantizers.223 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.69e-01 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.gate_up_proj_weight_quantizers.224 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.13e-01 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.gate_up_proj_weight_quantizers.225 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.23e-01 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.gate_up_proj_weight_quantizers.226 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.gate_up_proj_weight_quantizers.227 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.40e-01 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.gate_up_proj_weight_quantizers.228 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.50e-01 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.gate_up_proj_weight_quantizers.229 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.21e-01 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.gate_up_proj_weight_quantizers.230 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.gate_up_proj_weight_quantizers.231 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.12e-01 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.gate_up_proj_weight_quantizers.232 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.gate_up_proj_weight_quantizers.233 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.18e-01 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.gate_up_proj_weight_quantizers.234 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.gate_up_proj_weight_quantizers.235 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.22e-01 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.gate_up_proj_weight_quantizers.236 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.17e-01 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.gate_up_proj_weight_quantizers.237 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.72e-01 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.gate_up_proj_weight_quantizers.238 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.gate_up_proj_weight_quantizers.239 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.gate_up_proj_weight_quantizers.240 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.gate_up_proj_weight_quantizers.241 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.58e-01 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.gate_up_proj_weight_quantizers.242 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.gate_up_proj_weight_quantizers.243 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.gate_up_proj_weight_quantizers.244 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.gate_up_proj_weight_quantizers.245 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.gate_up_proj_weight_quantizers.246 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.12e-01 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.gate_up_proj_weight_quantizers.247 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.gate_up_proj_weight_quantizers.248 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.gate_up_proj_weight_quantizers.249 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.56e-01 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.gate_up_proj_weight_quantizers.250 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.gate_up_proj_weight_quantizers.251 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.77e-01 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.gate_up_proj_weight_quantizers.252 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.24e-01 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.gate_up_proj_weight_quantizers.253 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.30e-01 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.gate_up_proj_weight_quantizers.254 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.11e-01 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.gate_up_proj_weight_quantizers.255 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.down_proj_input_quantizer TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.89e+01 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.down_proj_weight_quantizers.0 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.down_proj_weight_quantizers.1 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.down_proj_weight_quantizers.2 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.down_proj_weight_quantizers.3 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.down_proj_weight_quantizers.4 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.down_proj_weight_quantizers.5 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.37e-01 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.down_proj_weight_quantizers.6 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.54e-02 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.down_proj_weight_quantizers.7 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.down_proj_weight_quantizers.8 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.down_proj_weight_quantizers.9 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.40e-02 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.down_proj_weight_quantizers.10 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.down_proj_weight_quantizers.11 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.13e-01 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.down_proj_weight_quantizers.12 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.54e-02 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.down_proj_weight_quantizers.13 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.down_proj_weight_quantizers.14 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.down_proj_weight_quantizers.15 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.down_proj_weight_quantizers.16 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.64e-02 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.down_proj_weight_quantizers.17 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.down_proj_weight_quantizers.18 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.down_proj_weight_quantizers.19 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.13e-01 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.down_proj_weight_quantizers.20 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.down_proj_weight_quantizers.21 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.56e-01 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.down_proj_weight_quantizers.22 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.59e-02 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.down_proj_weight_quantizers.23 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.down_proj_weight_quantizers.24 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.down_proj_weight_quantizers.25 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.down_proj_weight_quantizers.26 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.down_proj_weight_quantizers.27 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.down_proj_weight_quantizers.28 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.50e-02 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.down_proj_weight_quantizers.29 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.down_proj_weight_quantizers.30 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.down_proj_weight_quantizers.31 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.59e-02 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.down_proj_weight_quantizers.32 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.down_proj_weight_quantizers.33 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.54e-02 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.down_proj_weight_quantizers.34 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.down_proj_weight_quantizers.35 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.59e-01 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.down_proj_weight_quantizers.36 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.down_proj_weight_quantizers.37 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.down_proj_weight_quantizers.38 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.down_proj_weight_quantizers.39 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.down_proj_weight_quantizers.40 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.down_proj_weight_quantizers.41 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.down_proj_weight_quantizers.42 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.down_proj_weight_quantizers.43 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.down_proj_weight_quantizers.44 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.down_proj_weight_quantizers.45 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.down_proj_weight_quantizers.46 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.down_proj_weight_quantizers.47 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.down_proj_weight_quantizers.48 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.down_proj_weight_quantizers.49 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.34e-01 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.down_proj_weight_quantizers.50 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.down_proj_weight_quantizers.51 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.down_proj_weight_quantizers.52 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.down_proj_weight_quantizers.53 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.down_proj_weight_quantizers.54 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.35e-02 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.down_proj_weight_quantizers.55 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.down_proj_weight_quantizers.56 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.down_proj_weight_quantizers.57 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.59e-02 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.down_proj_weight_quantizers.58 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.down_proj_weight_quantizers.59 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.down_proj_weight_quantizers.60 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.down_proj_weight_quantizers.61 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.down_proj_weight_quantizers.62 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.down_proj_weight_quantizers.63 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.down_proj_weight_quantizers.64 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.down_proj_weight_quantizers.65 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.down_proj_weight_quantizers.66 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.down_proj_weight_quantizers.67 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.down_proj_weight_quantizers.68 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.74e-01 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.down_proj_weight_quantizers.69 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.down_proj_weight_quantizers.70 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.down_proj_weight_quantizers.71 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.16e-01 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.down_proj_weight_quantizers.72 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.down_proj_weight_quantizers.73 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.down_proj_weight_quantizers.74 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.50e-01 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.down_proj_weight_quantizers.75 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.down_proj_weight_quantizers.76 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.down_proj_weight_quantizers.77 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.45e-02 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.down_proj_weight_quantizers.78 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.down_proj_weight_quantizers.79 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.down_proj_weight_quantizers.80 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.down_proj_weight_quantizers.81 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.down_proj_weight_quantizers.82 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.down_proj_weight_quantizers.83 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.64e-02 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.down_proj_weight_quantizers.84 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.down_proj_weight_quantizers.85 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.down_proj_weight_quantizers.86 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.71e-01 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.down_proj_weight_quantizers.87 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.down_proj_weight_quantizers.88 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.68e-01 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.down_proj_weight_quantizers.89 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.down_proj_weight_quantizers.90 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.down_proj_weight_quantizers.91 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.down_proj_weight_quantizers.92 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.down_proj_weight_quantizers.93 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.down_proj_weight_quantizers.94 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.down_proj_weight_quantizers.95 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.down_proj_weight_quantizers.96 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.36e-01 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.down_proj_weight_quantizers.97 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.down_proj_weight_quantizers.98 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.down_proj_weight_quantizers.99 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.down_proj_weight_quantizers.100 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.86e-01 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.down_proj_weight_quantizers.101 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.down_proj_weight_quantizers.102 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.down_proj_weight_quantizers.103 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.45e-02 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.down_proj_weight_quantizers.104 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.25e-01 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.down_proj_weight_quantizers.105 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.down_proj_weight_quantizers.106 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.down_proj_weight_quantizers.107 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.64e-02 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.down_proj_weight_quantizers.108 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.down_proj_weight_quantizers.109 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.down_proj_weight_quantizers.110 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.down_proj_weight_quantizers.111 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.64e-02 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.down_proj_weight_quantizers.112 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.50e-02 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.down_proj_weight_quantizers.113 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.down_proj_weight_quantizers.114 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.down_proj_weight_quantizers.115 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.down_proj_weight_quantizers.116 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.42e-01 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.down_proj_weight_quantizers.117 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.down_proj_weight_quantizers.118 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.down_proj_weight_quantizers.119 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.down_proj_weight_quantizers.120 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.down_proj_weight_quantizers.121 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.down_proj_weight_quantizers.122 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.down_proj_weight_quantizers.123 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.76e-01 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.down_proj_weight_quantizers.124 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.down_proj_weight_quantizers.125 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.down_proj_weight_quantizers.126 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.down_proj_weight_quantizers.127 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.down_proj_weight_quantizers.128 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.down_proj_weight_quantizers.129 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.down_proj_weight_quantizers.130 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.68e-01 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.down_proj_weight_quantizers.131 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.down_proj_weight_quantizers.132 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.down_proj_weight_quantizers.133 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.down_proj_weight_quantizers.134 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.down_proj_weight_quantizers.135 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.down_proj_weight_quantizers.136 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.30e-02 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.down_proj_weight_quantizers.137 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.down_proj_weight_quantizers.138 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.87e-01 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.down_proj_weight_quantizers.139 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.down_proj_weight_quantizers.140 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.down_proj_weight_quantizers.141 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.down_proj_weight_quantizers.142 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.18e-01 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.down_proj_weight_quantizers.143 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.down_proj_weight_quantizers.144 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.down_proj_weight_quantizers.145 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.down_proj_weight_quantizers.146 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.50e-01 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.down_proj_weight_quantizers.147 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.down_proj_weight_quantizers.148 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.60e-01 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.down_proj_weight_quantizers.149 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.54e-02 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.down_proj_weight_quantizers.150 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.down_proj_weight_quantizers.151 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.down_proj_weight_quantizers.152 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.down_proj_weight_quantizers.153 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.down_proj_weight_quantizers.154 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.down_proj_weight_quantizers.155 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.down_proj_weight_quantizers.156 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.down_proj_weight_quantizers.157 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.down_proj_weight_quantizers.158 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.down_proj_weight_quantizers.159 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.down_proj_weight_quantizers.160 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.down_proj_weight_quantizers.161 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.59e-02 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.down_proj_weight_quantizers.162 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.14e-01 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.down_proj_weight_quantizers.163 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.45e-02 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.down_proj_weight_quantizers.164 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.down_proj_weight_quantizers.165 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.down_proj_weight_quantizers.166 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.down_proj_weight_quantizers.167 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.down_proj_weight_quantizers.168 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.down_proj_weight_quantizers.169 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.down_proj_weight_quantizers.170 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.down_proj_weight_quantizers.171 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.down_proj_weight_quantizers.172 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.down_proj_weight_quantizers.173 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.down_proj_weight_quantizers.174 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.down_proj_weight_quantizers.175 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.down_proj_weight_quantizers.176 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.down_proj_weight_quantizers.177 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.42e-01 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.down_proj_weight_quantizers.178 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.down_proj_weight_quantizers.179 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.down_proj_weight_quantizers.180 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.down_proj_weight_quantizers.181 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.down_proj_weight_quantizers.182 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.down_proj_weight_quantizers.183 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.down_proj_weight_quantizers.184 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.down_proj_weight_quantizers.185 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.down_proj_weight_quantizers.186 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.22e-01 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.down_proj_weight_quantizers.187 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.down_proj_weight_quantizers.188 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.down_proj_weight_quantizers.189 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.down_proj_weight_quantizers.190 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.down_proj_weight_quantizers.191 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.down_proj_weight_quantizers.192 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.down_proj_weight_quantizers.193 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.64e-02 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.down_proj_weight_quantizers.194 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.down_proj_weight_quantizers.195 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.64e-02 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.down_proj_weight_quantizers.196 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.down_proj_weight_quantizers.197 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.down_proj_weight_quantizers.198 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.down_proj_weight_quantizers.199 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.down_proj_weight_quantizers.200 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.down_proj_weight_quantizers.201 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.down_proj_weight_quantizers.202 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.down_proj_weight_quantizers.203 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.59e-02 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.down_proj_weight_quantizers.204 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.down_proj_weight_quantizers.205 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.down_proj_weight_quantizers.206 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.54e-02 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.down_proj_weight_quantizers.207 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.down_proj_weight_quantizers.208 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.49e-01 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.down_proj_weight_quantizers.209 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.down_proj_weight_quantizers.210 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.59e-02 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.down_proj_weight_quantizers.211 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.down_proj_weight_quantizers.212 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.63e-01 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.down_proj_weight_quantizers.213 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.down_proj_weight_quantizers.214 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.down_proj_weight_quantizers.215 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.35e-02 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.down_proj_weight_quantizers.216 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.down_proj_weight_quantizers.217 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.down_proj_weight_quantizers.218 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.down_proj_weight_quantizers.219 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.26e-01 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.down_proj_weight_quantizers.220 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.down_proj_weight_quantizers.221 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.40e-01 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.down_proj_weight_quantizers.222 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.87e-01 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.down_proj_weight_quantizers.223 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.down_proj_weight_quantizers.224 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.64e-02 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.down_proj_weight_quantizers.225 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.down_proj_weight_quantizers.226 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.down_proj_weight_quantizers.227 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.down_proj_weight_quantizers.228 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.27e-01 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.down_proj_weight_quantizers.229 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.42e-01 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.down_proj_weight_quantizers.230 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.down_proj_weight_quantizers.231 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.19e-01 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.down_proj_weight_quantizers.232 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.29e-01 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.down_proj_weight_quantizers.233 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.59e-02 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.down_proj_weight_quantizers.234 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.down_proj_weight_quantizers.235 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.down_proj_weight_quantizers.236 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.down_proj_weight_quantizers.237 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.86e-01 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.down_proj_weight_quantizers.238 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.down_proj_weight_quantizers.239 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.down_proj_weight_quantizers.240 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.down_proj_weight_quantizers.241 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.69e-01 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.down_proj_weight_quantizers.242 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.down_proj_weight_quantizers.243 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.41e-01 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.down_proj_weight_quantizers.244 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.down_proj_weight_quantizers.245 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.down_proj_weight_quantizers.246 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.down_proj_weight_quantizers.247 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.down_proj_weight_quantizers.248 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.down_proj_weight_quantizers.249 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.down_proj_weight_quantizers.250 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.46e-01 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.down_proj_weight_quantizers.251 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.down_proj_weight_quantizers.252 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.down_proj_weight_quantizers.253 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.34e-01 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.down_proj_weight_quantizers.254 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.91e-01 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.down_proj_weight_quantizers.255 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.42.mlp.shared_experts.gate_proj.input_quantizer TensorQuantizer(disabled) +model.layers.42.mlp.shared_experts.gate_proj.output_quantizer TensorQuantizer(disabled) +model.layers.42.mlp.shared_experts.gate_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.42.mlp.shared_experts.up_proj.input_quantizer TensorQuantizer(disabled) +model.layers.42.mlp.shared_experts.up_proj.output_quantizer TensorQuantizer(disabled) +model.layers.42.mlp.shared_experts.up_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.42.mlp.shared_experts.down_proj.input_quantizer TensorQuantizer(disabled) +model.layers.42.mlp.shared_experts.down_proj.output_quantizer TensorQuantizer(disabled) +model.layers.42.mlp.shared_experts.down_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.43.self_attn.q_a_proj.input_quantizer TensorQuantizer(disabled) +model.layers.43.self_attn.q_a_proj.output_quantizer TensorQuantizer(disabled) +model.layers.43.self_attn.q_a_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.43.self_attn.q_b_proj.input_quantizer TensorQuantizer(disabled) +model.layers.43.self_attn.q_b_proj.output_quantizer TensorQuantizer(disabled) +model.layers.43.self_attn.q_b_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.43.self_attn.kv_a_proj_with_mqa.input_quantizer TensorQuantizer(disabled) +model.layers.43.self_attn.kv_a_proj_with_mqa.output_quantizer TensorQuantizer(disabled) +model.layers.43.self_attn.kv_a_proj_with_mqa.weight_quantizer TensorQuantizer(disabled) +model.layers.43.self_attn.kv_b_proj.input_quantizer TensorQuantizer(disabled) +model.layers.43.self_attn.kv_b_proj.output_quantizer TensorQuantizer(disabled) +model.layers.43.self_attn.kv_b_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.43.self_attn.o_proj.input_quantizer TensorQuantizer(disabled) +model.layers.43.self_attn.o_proj.output_quantizer TensorQuantizer(disabled) +model.layers.43.self_attn.o_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.43.self_attn.q_bmm_quantizer TensorQuantizer(disabled) +model.layers.43.self_attn.k_bmm_quantizer TensorQuantizer((4, 3) bit fake per-tensor amax=4.48e+02(const) calibrator=MaxCalibrator quant) +model.layers.43.self_attn.v_bmm_quantizer TensorQuantizer((4, 3) bit fake per-tensor amax=4.48e+02(const) calibrator=MaxCalibrator quant) +model.layers.43.self_attn.p_bmm_quantizer TensorQuantizer(disabled) +model.layers.43.mlp.experts.gate_up_proj_input_quantizer TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.02e+00 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.gate_up_proj_weight_quantizers.0 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.gate_up_proj_weight_quantizers.1 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.17e-01 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.gate_up_proj_weight_quantizers.2 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.gate_up_proj_weight_quantizers.3 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.gate_up_proj_weight_quantizers.4 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.gate_up_proj_weight_quantizers.5 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.gate_up_proj_weight_quantizers.6 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.gate_up_proj_weight_quantizers.7 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.gate_up_proj_weight_quantizers.8 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.gate_up_proj_weight_quantizers.9 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.gate_up_proj_weight_quantizers.10 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.gate_up_proj_weight_quantizers.11 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.gate_up_proj_weight_quantizers.12 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.34e-01 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.gate_up_proj_weight_quantizers.13 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.gate_up_proj_weight_quantizers.14 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.gate_up_proj_weight_quantizers.15 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.gate_up_proj_weight_quantizers.16 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.gate_up_proj_weight_quantizers.17 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.gate_up_proj_weight_quantizers.18 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=4.43e-01 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.gate_up_proj_weight_quantizers.19 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.gate_up_proj_weight_quantizers.20 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.gate_up_proj_weight_quantizers.21 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.gate_up_proj_weight_quantizers.22 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.29e-01 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.gate_up_proj_weight_quantizers.23 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.50e-01 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.gate_up_proj_weight_quantizers.24 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.34e-01 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.gate_up_proj_weight_quantizers.25 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.gate_up_proj_weight_quantizers.26 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.19e-01 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.gate_up_proj_weight_quantizers.27 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.gate_up_proj_weight_quantizers.28 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.32e-01 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.gate_up_proj_weight_quantizers.29 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.gate_up_proj_weight_quantizers.30 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.gate_up_proj_weight_quantizers.31 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.81e-01 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.gate_up_proj_weight_quantizers.32 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.gate_up_proj_weight_quantizers.33 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.gate_up_proj_weight_quantizers.34 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.gate_up_proj_weight_quantizers.35 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.gate_up_proj_weight_quantizers.36 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.gate_up_proj_weight_quantizers.37 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.33e-01 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.gate_up_proj_weight_quantizers.38 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.43e-01 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.gate_up_proj_weight_quantizers.39 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.gate_up_proj_weight_quantizers.40 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.63e-01 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.gate_up_proj_weight_quantizers.41 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.gate_up_proj_weight_quantizers.42 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.gate_up_proj_weight_quantizers.43 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.27e-01 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.gate_up_proj_weight_quantizers.44 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.gate_up_proj_weight_quantizers.45 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.67e-01 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.gate_up_proj_weight_quantizers.46 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.35e-01 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.gate_up_proj_weight_quantizers.47 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.gate_up_proj_weight_quantizers.48 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.gate_up_proj_weight_quantizers.49 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.28e-01 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.gate_up_proj_weight_quantizers.50 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.gate_up_proj_weight_quantizers.51 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.gate_up_proj_weight_quantizers.52 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.gate_up_proj_weight_quantizers.53 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.gate_up_proj_weight_quantizers.54 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.gate_up_proj_weight_quantizers.55 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.28e-01 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.gate_up_proj_weight_quantizers.56 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.22e-01 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.gate_up_proj_weight_quantizers.57 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.34e-01 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.gate_up_proj_weight_quantizers.58 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.67e-01 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.gate_up_proj_weight_quantizers.59 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.38e-01 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.gate_up_proj_weight_quantizers.60 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.gate_up_proj_weight_quantizers.61 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.14e-01 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.gate_up_proj_weight_quantizers.62 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.32e-01 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.gate_up_proj_weight_quantizers.63 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.25e-01 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.gate_up_proj_weight_quantizers.64 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.gate_up_proj_weight_quantizers.65 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.38e-01 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.gate_up_proj_weight_quantizers.66 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.gate_up_proj_weight_quantizers.67 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.gate_up_proj_weight_quantizers.68 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.37e-01 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.gate_up_proj_weight_quantizers.69 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.81e-01 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.gate_up_proj_weight_quantizers.70 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.gate_up_proj_weight_quantizers.71 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.gate_up_proj_weight_quantizers.72 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.gate_up_proj_weight_quantizers.73 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=4.90e-01 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.gate_up_proj_weight_quantizers.74 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.gate_up_proj_weight_quantizers.75 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.gate_up_proj_weight_quantizers.76 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.gate_up_proj_weight_quantizers.77 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.gate_up_proj_weight_quantizers.78 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.gate_up_proj_weight_quantizers.79 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.gate_up_proj_weight_quantizers.80 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.gate_up_proj_weight_quantizers.81 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.27e-01 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.gate_up_proj_weight_quantizers.82 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.gate_up_proj_weight_quantizers.83 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.gate_up_proj_weight_quantizers.84 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.14e-01 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.gate_up_proj_weight_quantizers.85 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.gate_up_proj_weight_quantizers.86 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.gate_up_proj_weight_quantizers.87 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.88e-01 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.gate_up_proj_weight_quantizers.88 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.17e-01 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.gate_up_proj_weight_quantizers.89 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.62e-01 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.gate_up_proj_weight_quantizers.90 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.54e-01 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.gate_up_proj_weight_quantizers.91 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.gate_up_proj_weight_quantizers.92 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.18e-01 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.gate_up_proj_weight_quantizers.93 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.29e-01 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.gate_up_proj_weight_quantizers.94 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.gate_up_proj_weight_quantizers.95 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.35e-01 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.gate_up_proj_weight_quantizers.96 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.gate_up_proj_weight_quantizers.97 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.gate_up_proj_weight_quantizers.98 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.gate_up_proj_weight_quantizers.99 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.18e-01 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.gate_up_proj_weight_quantizers.100 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.15e-01 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.gate_up_proj_weight_quantizers.101 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.66e-01 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.gate_up_proj_weight_quantizers.102 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.28e-01 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.gate_up_proj_weight_quantizers.103 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.gate_up_proj_weight_quantizers.104 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.27e-01 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.gate_up_proj_weight_quantizers.105 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.29e-01 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.gate_up_proj_weight_quantizers.106 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.gate_up_proj_weight_quantizers.107 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.69e-01 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.gate_up_proj_weight_quantizers.108 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.gate_up_proj_weight_quantizers.109 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.39e-01 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.gate_up_proj_weight_quantizers.110 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.55e-01 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.gate_up_proj_weight_quantizers.111 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.gate_up_proj_weight_quantizers.112 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.72e-01 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.gate_up_proj_weight_quantizers.113 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.gate_up_proj_weight_quantizers.114 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.46e-01 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.gate_up_proj_weight_quantizers.115 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.gate_up_proj_weight_quantizers.116 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.32e-01 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.gate_up_proj_weight_quantizers.117 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.gate_up_proj_weight_quantizers.118 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.gate_up_proj_weight_quantizers.119 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.gate_up_proj_weight_quantizers.120 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.85e-01 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.gate_up_proj_weight_quantizers.121 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.gate_up_proj_weight_quantizers.122 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.14e-01 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.gate_up_proj_weight_quantizers.123 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.34e-01 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.gate_up_proj_weight_quantizers.124 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.gate_up_proj_weight_quantizers.125 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.gate_up_proj_weight_quantizers.126 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.40e-01 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.gate_up_proj_weight_quantizers.127 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.gate_up_proj_weight_quantizers.128 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.gate_up_proj_weight_quantizers.129 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.gate_up_proj_weight_quantizers.130 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.gate_up_proj_weight_quantizers.131 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.62e-01 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.gate_up_proj_weight_quantizers.132 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.gate_up_proj_weight_quantizers.133 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.gate_up_proj_weight_quantizers.134 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.14e-01 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.gate_up_proj_weight_quantizers.135 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.gate_up_proj_weight_quantizers.136 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.20e-01 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.gate_up_proj_weight_quantizers.137 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.27e-01 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.gate_up_proj_weight_quantizers.138 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.gate_up_proj_weight_quantizers.139 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.91e-01 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.gate_up_proj_weight_quantizers.140 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.gate_up_proj_weight_quantizers.141 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.24e-01 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.gate_up_proj_weight_quantizers.142 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.46e-01 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.gate_up_proj_weight_quantizers.143 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.gate_up_proj_weight_quantizers.144 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.28e-01 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.gate_up_proj_weight_quantizers.145 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.28e-01 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.gate_up_proj_weight_quantizers.146 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.gate_up_proj_weight_quantizers.147 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.70e-01 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.gate_up_proj_weight_quantizers.148 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.16e-01 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.gate_up_proj_weight_quantizers.149 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.27e-01 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.gate_up_proj_weight_quantizers.150 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.52e-01 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.gate_up_proj_weight_quantizers.151 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.gate_up_proj_weight_quantizers.152 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.gate_up_proj_weight_quantizers.153 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.gate_up_proj_weight_quantizers.154 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.gate_up_proj_weight_quantizers.155 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.gate_up_proj_weight_quantizers.156 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.47e-01 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.gate_up_proj_weight_quantizers.157 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.gate_up_proj_weight_quantizers.158 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.gate_up_proj_weight_quantizers.159 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.60e-01 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.gate_up_proj_weight_quantizers.160 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.46e-01 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.gate_up_proj_weight_quantizers.161 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.gate_up_proj_weight_quantizers.162 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.63e-01 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.gate_up_proj_weight_quantizers.163 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.25e-01 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.gate_up_proj_weight_quantizers.164 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.gate_up_proj_weight_quantizers.165 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.gate_up_proj_weight_quantizers.166 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.47e-01 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.gate_up_proj_weight_quantizers.167 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.gate_up_proj_weight_quantizers.168 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.gate_up_proj_weight_quantizers.169 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.23e-01 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.gate_up_proj_weight_quantizers.170 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.gate_up_proj_weight_quantizers.171 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.30e-01 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.gate_up_proj_weight_quantizers.172 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.gate_up_proj_weight_quantizers.173 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.22e-01 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.gate_up_proj_weight_quantizers.174 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.15e-01 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.gate_up_proj_weight_quantizers.175 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.gate_up_proj_weight_quantizers.176 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.32e-01 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.gate_up_proj_weight_quantizers.177 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.47e-01 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.gate_up_proj_weight_quantizers.178 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.23e-01 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.gate_up_proj_weight_quantizers.179 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.gate_up_proj_weight_quantizers.180 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.16e-01 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.gate_up_proj_weight_quantizers.181 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.gate_up_proj_weight_quantizers.182 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.30e-01 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.gate_up_proj_weight_quantizers.183 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.gate_up_proj_weight_quantizers.184 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.gate_up_proj_weight_quantizers.185 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.gate_up_proj_weight_quantizers.186 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.gate_up_proj_weight_quantizers.187 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.31e-01 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.gate_up_proj_weight_quantizers.188 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.gate_up_proj_weight_quantizers.189 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.gate_up_proj_weight_quantizers.190 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.gate_up_proj_weight_quantizers.191 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.35e-01 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.gate_up_proj_weight_quantizers.192 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.gate_up_proj_weight_quantizers.193 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.gate_up_proj_weight_quantizers.194 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.19e-01 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.gate_up_proj_weight_quantizers.195 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.44e-01 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.gate_up_proj_weight_quantizers.196 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.gate_up_proj_weight_quantizers.197 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.gate_up_proj_weight_quantizers.198 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.gate_up_proj_weight_quantizers.199 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.gate_up_proj_weight_quantizers.200 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.gate_up_proj_weight_quantizers.201 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.gate_up_proj_weight_quantizers.202 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.gate_up_proj_weight_quantizers.203 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.gate_up_proj_weight_quantizers.204 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.18e-01 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.gate_up_proj_weight_quantizers.205 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.gate_up_proj_weight_quantizers.206 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.26e-01 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.gate_up_proj_weight_quantizers.207 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.gate_up_proj_weight_quantizers.208 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.80e-01 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.gate_up_proj_weight_quantizers.209 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.33e-01 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.gate_up_proj_weight_quantizers.210 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.23e-01 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.gate_up_proj_weight_quantizers.211 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.26e-01 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.gate_up_proj_weight_quantizers.212 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.gate_up_proj_weight_quantizers.213 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.01e-01 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.gate_up_proj_weight_quantizers.214 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.gate_up_proj_weight_quantizers.215 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.22e-01 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.gate_up_proj_weight_quantizers.216 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.15e-01 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.gate_up_proj_weight_quantizers.217 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.gate_up_proj_weight_quantizers.218 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.gate_up_proj_weight_quantizers.219 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.gate_up_proj_weight_quantizers.220 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.gate_up_proj_weight_quantizers.221 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.gate_up_proj_weight_quantizers.222 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.16e-01 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.gate_up_proj_weight_quantizers.223 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.gate_up_proj_weight_quantizers.224 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.gate_up_proj_weight_quantizers.225 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.22e-01 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.gate_up_proj_weight_quantizers.226 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.gate_up_proj_weight_quantizers.227 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.gate_up_proj_weight_quantizers.228 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.27e-01 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.gate_up_proj_weight_quantizers.229 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.44e-01 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.gate_up_proj_weight_quantizers.230 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.99e-01 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.gate_up_proj_weight_quantizers.231 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.gate_up_proj_weight_quantizers.232 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.gate_up_proj_weight_quantizers.233 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.46e-01 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.gate_up_proj_weight_quantizers.234 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.23e-01 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.gate_up_proj_weight_quantizers.235 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.gate_up_proj_weight_quantizers.236 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.gate_up_proj_weight_quantizers.237 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.gate_up_proj_weight_quantizers.238 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.03e-01 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.gate_up_proj_weight_quantizers.239 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.27e-01 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.gate_up_proj_weight_quantizers.240 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.gate_up_proj_weight_quantizers.241 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.37e-01 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.gate_up_proj_weight_quantizers.242 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.14e-01 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.gate_up_proj_weight_quantizers.243 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.13e-01 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.gate_up_proj_weight_quantizers.244 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.53e-01 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.gate_up_proj_weight_quantizers.245 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.gate_up_proj_weight_quantizers.246 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.gate_up_proj_weight_quantizers.247 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.14e-01 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.gate_up_proj_weight_quantizers.248 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.gate_up_proj_weight_quantizers.249 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.gate_up_proj_weight_quantizers.250 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.52e-01 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.gate_up_proj_weight_quantizers.251 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.gate_up_proj_weight_quantizers.252 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.27e-01 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.gate_up_proj_weight_quantizers.253 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.46e-01 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.gate_up_proj_weight_quantizers.254 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.gate_up_proj_weight_quantizers.255 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.29e-01 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.down_proj_input_quantizer TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.91e+01 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.down_proj_weight_quantizers.0 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.down_proj_weight_quantizers.1 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.99e-01 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.down_proj_weight_quantizers.2 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.down_proj_weight_quantizers.3 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.down_proj_weight_quantizers.4 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.down_proj_weight_quantizers.5 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.down_proj_weight_quantizers.6 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.down_proj_weight_quantizers.7 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.down_proj_weight_quantizers.8 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.down_proj_weight_quantizers.9 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.down_proj_weight_quantizers.10 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.down_proj_weight_quantizers.11 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.down_proj_weight_quantizers.12 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.42e-01 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.down_proj_weight_quantizers.13 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.down_proj_weight_quantizers.14 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.down_proj_weight_quantizers.15 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.down_proj_weight_quantizers.16 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.down_proj_weight_quantizers.17 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.down_proj_weight_quantizers.18 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.90e-01 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.down_proj_weight_quantizers.19 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.64e-02 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.down_proj_weight_quantizers.20 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.down_proj_weight_quantizers.21 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.down_proj_weight_quantizers.22 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.45e-02 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.down_proj_weight_quantizers.23 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.down_proj_weight_quantizers.24 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.15e-02 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.down_proj_weight_quantizers.25 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.down_proj_weight_quantizers.26 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.down_proj_weight_quantizers.27 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.down_proj_weight_quantizers.28 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.19e-01 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.down_proj_weight_quantizers.29 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.59e-02 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.down_proj_weight_quantizers.30 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.down_proj_weight_quantizers.31 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.89e-01 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.down_proj_weight_quantizers.32 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.down_proj_weight_quantizers.33 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.down_proj_weight_quantizers.34 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.64e-02 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.down_proj_weight_quantizers.35 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.50e-02 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.down_proj_weight_quantizers.36 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.down_proj_weight_quantizers.37 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.down_proj_weight_quantizers.38 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.65e-01 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.down_proj_weight_quantizers.39 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.down_proj_weight_quantizers.40 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.64e-02 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.down_proj_weight_quantizers.41 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.down_proj_weight_quantizers.42 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.down_proj_weight_quantizers.43 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.down_proj_weight_quantizers.44 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.down_proj_weight_quantizers.45 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.70e-01 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.down_proj_weight_quantizers.46 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.down_proj_weight_quantizers.47 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.30e-01 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.down_proj_weight_quantizers.48 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.down_proj_weight_quantizers.49 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.down_proj_weight_quantizers.50 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.down_proj_weight_quantizers.51 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.46e-01 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.down_proj_weight_quantizers.52 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.down_proj_weight_quantizers.53 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.down_proj_weight_quantizers.54 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.down_proj_weight_quantizers.55 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.down_proj_weight_quantizers.56 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.45e-02 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.down_proj_weight_quantizers.57 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.down_proj_weight_quantizers.58 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.45e-02 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.down_proj_weight_quantizers.59 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.down_proj_weight_quantizers.60 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.down_proj_weight_quantizers.61 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.down_proj_weight_quantizers.62 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.94e-01 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.down_proj_weight_quantizers.63 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.down_proj_weight_quantizers.64 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.down_proj_weight_quantizers.65 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.down_proj_weight_quantizers.66 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.down_proj_weight_quantizers.67 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.down_proj_weight_quantizers.68 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.down_proj_weight_quantizers.69 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.down_proj_weight_quantizers.70 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.down_proj_weight_quantizers.71 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.down_proj_weight_quantizers.72 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.down_proj_weight_quantizers.73 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.36e-01 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.down_proj_weight_quantizers.74 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.down_proj_weight_quantizers.75 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.down_proj_weight_quantizers.76 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.50e-02 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.down_proj_weight_quantizers.77 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.down_proj_weight_quantizers.78 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.down_proj_weight_quantizers.79 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.down_proj_weight_quantizers.80 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.down_proj_weight_quantizers.81 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.down_proj_weight_quantizers.82 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.down_proj_weight_quantizers.83 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.down_proj_weight_quantizers.84 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.down_proj_weight_quantizers.85 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.down_proj_weight_quantizers.86 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.down_proj_weight_quantizers.87 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.56e-01 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.down_proj_weight_quantizers.88 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.down_proj_weight_quantizers.89 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.down_proj_weight_quantizers.90 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.down_proj_weight_quantizers.91 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.52e-01 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.down_proj_weight_quantizers.92 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.down_proj_weight_quantizers.93 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.down_proj_weight_quantizers.94 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.down_proj_weight_quantizers.95 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.down_proj_weight_quantizers.96 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.down_proj_weight_quantizers.97 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.down_proj_weight_quantizers.98 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.24e-01 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.down_proj_weight_quantizers.99 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.down_proj_weight_quantizers.100 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.down_proj_weight_quantizers.101 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.down_proj_weight_quantizers.102 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.down_proj_weight_quantizers.103 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.down_proj_weight_quantizers.104 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.down_proj_weight_quantizers.105 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.down_proj_weight_quantizers.106 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.54e-02 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.down_proj_weight_quantizers.107 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.69e-01 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.down_proj_weight_quantizers.108 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.down_proj_weight_quantizers.109 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.57e-01 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.down_proj_weight_quantizers.110 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.13e-01 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.down_proj_weight_quantizers.111 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.45e-02 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.down_proj_weight_quantizers.112 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.64e-02 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.down_proj_weight_quantizers.113 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.17e-01 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.down_proj_weight_quantizers.114 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.54e-02 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.down_proj_weight_quantizers.115 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.64e-02 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.down_proj_weight_quantizers.116 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.down_proj_weight_quantizers.117 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.down_proj_weight_quantizers.118 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.down_proj_weight_quantizers.119 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.down_proj_weight_quantizers.120 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.26e-01 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.down_proj_weight_quantizers.121 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.down_proj_weight_quantizers.122 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.down_proj_weight_quantizers.123 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.50e-02 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.down_proj_weight_quantizers.124 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.down_proj_weight_quantizers.125 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.down_proj_weight_quantizers.126 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.down_proj_weight_quantizers.127 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.down_proj_weight_quantizers.128 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.51e-01 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.down_proj_weight_quantizers.129 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.59e-02 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.down_proj_weight_quantizers.130 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.15e-01 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.down_proj_weight_quantizers.131 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.50e-02 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.down_proj_weight_quantizers.132 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.down_proj_weight_quantizers.133 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.down_proj_weight_quantizers.134 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.29e-01 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.down_proj_weight_quantizers.135 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.48e-01 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.down_proj_weight_quantizers.136 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.down_proj_weight_quantizers.137 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.down_proj_weight_quantizers.138 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.down_proj_weight_quantizers.139 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.down_proj_weight_quantizers.140 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.down_proj_weight_quantizers.141 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.45e-02 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.down_proj_weight_quantizers.142 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.01e-01 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.down_proj_weight_quantizers.143 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.down_proj_weight_quantizers.144 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.31e-01 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.down_proj_weight_quantizers.145 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.23e-01 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.down_proj_weight_quantizers.146 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.down_proj_weight_quantizers.147 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.down_proj_weight_quantizers.148 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.down_proj_weight_quantizers.149 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.down_proj_weight_quantizers.150 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.45e-02 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.down_proj_weight_quantizers.151 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.down_proj_weight_quantizers.152 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.down_proj_weight_quantizers.153 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.down_proj_weight_quantizers.154 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.down_proj_weight_quantizers.155 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.down_proj_weight_quantizers.156 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.14e-01 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.down_proj_weight_quantizers.157 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.59e-02 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.down_proj_weight_quantizers.158 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.down_proj_weight_quantizers.159 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.down_proj_weight_quantizers.160 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.59e-02 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.down_proj_weight_quantizers.161 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.down_proj_weight_quantizers.162 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.down_proj_weight_quantizers.163 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.down_proj_weight_quantizers.164 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.54e-02 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.down_proj_weight_quantizers.165 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.54e-01 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.down_proj_weight_quantizers.166 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.50e-02 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.down_proj_weight_quantizers.167 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.down_proj_weight_quantizers.168 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.down_proj_weight_quantizers.169 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.down_proj_weight_quantizers.170 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.down_proj_weight_quantizers.171 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.down_proj_weight_quantizers.172 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.down_proj_weight_quantizers.173 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.54e-02 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.down_proj_weight_quantizers.174 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.down_proj_weight_quantizers.175 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.45e-02 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.down_proj_weight_quantizers.176 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.18e-01 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.down_proj_weight_quantizers.177 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.16e-01 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.down_proj_weight_quantizers.178 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.down_proj_weight_quantizers.179 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.down_proj_weight_quantizers.180 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.down_proj_weight_quantizers.181 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.64e-02 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.down_proj_weight_quantizers.182 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.down_proj_weight_quantizers.183 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.down_proj_weight_quantizers.184 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.down_proj_weight_quantizers.185 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.down_proj_weight_quantizers.186 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.50e-02 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.down_proj_weight_quantizers.187 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.89e-01 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.down_proj_weight_quantizers.188 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.down_proj_weight_quantizers.189 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.down_proj_weight_quantizers.190 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.down_proj_weight_quantizers.191 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.26e-01 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.down_proj_weight_quantizers.192 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.down_proj_weight_quantizers.193 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.down_proj_weight_quantizers.194 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.down_proj_weight_quantizers.195 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.down_proj_weight_quantizers.196 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.down_proj_weight_quantizers.197 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.down_proj_weight_quantizers.198 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.down_proj_weight_quantizers.199 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.down_proj_weight_quantizers.200 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.down_proj_weight_quantizers.201 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.down_proj_weight_quantizers.202 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.down_proj_weight_quantizers.203 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.down_proj_weight_quantizers.204 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.down_proj_weight_quantizers.205 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.down_proj_weight_quantizers.206 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.16e-01 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.down_proj_weight_quantizers.207 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.down_proj_weight_quantizers.208 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.74e-01 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.down_proj_weight_quantizers.209 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.29e-01 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.down_proj_weight_quantizers.210 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.50e-02 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.down_proj_weight_quantizers.211 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.24e-01 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.down_proj_weight_quantizers.212 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.down_proj_weight_quantizers.213 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.down_proj_weight_quantizers.214 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.27e-01 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.down_proj_weight_quantizers.215 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.down_proj_weight_quantizers.216 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.down_proj_weight_quantizers.217 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.59e-02 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.down_proj_weight_quantizers.218 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.down_proj_weight_quantizers.219 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.64e-02 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.down_proj_weight_quantizers.220 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.down_proj_weight_quantizers.221 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.down_proj_weight_quantizers.222 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.down_proj_weight_quantizers.223 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.down_proj_weight_quantizers.224 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.down_proj_weight_quantizers.225 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.down_proj_weight_quantizers.226 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.54e-02 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.down_proj_weight_quantizers.227 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.64e-02 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.down_proj_weight_quantizers.228 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.down_proj_weight_quantizers.229 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.87e-01 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.down_proj_weight_quantizers.230 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.18e-01 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.down_proj_weight_quantizers.231 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.down_proj_weight_quantizers.232 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.down_proj_weight_quantizers.233 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.91e-01 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.down_proj_weight_quantizers.234 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.down_proj_weight_quantizers.235 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.down_proj_weight_quantizers.236 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.down_proj_weight_quantizers.237 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.down_proj_weight_quantizers.238 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.61e-01 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.down_proj_weight_quantizers.239 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.down_proj_weight_quantizers.240 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.down_proj_weight_quantizers.241 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.35e-02 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.down_proj_weight_quantizers.242 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.down_proj_weight_quantizers.243 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.down_proj_weight_quantizers.244 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.down_proj_weight_quantizers.245 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.down_proj_weight_quantizers.246 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.50e-02 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.down_proj_weight_quantizers.247 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.down_proj_weight_quantizers.248 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.down_proj_weight_quantizers.249 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.down_proj_weight_quantizers.250 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.down_proj_weight_quantizers.251 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.down_proj_weight_quantizers.252 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.down_proj_weight_quantizers.253 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.down_proj_weight_quantizers.254 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.down_proj_weight_quantizers.255 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.52e-01 calibrator=MaxCalibrator quant) +model.layers.43.mlp.shared_experts.gate_proj.input_quantizer TensorQuantizer(disabled) +model.layers.43.mlp.shared_experts.gate_proj.output_quantizer TensorQuantizer(disabled) +model.layers.43.mlp.shared_experts.gate_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.43.mlp.shared_experts.up_proj.input_quantizer TensorQuantizer(disabled) +model.layers.43.mlp.shared_experts.up_proj.output_quantizer TensorQuantizer(disabled) +model.layers.43.mlp.shared_experts.up_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.43.mlp.shared_experts.down_proj.input_quantizer TensorQuantizer(disabled) +model.layers.43.mlp.shared_experts.down_proj.output_quantizer TensorQuantizer(disabled) +model.layers.43.mlp.shared_experts.down_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.44.self_attn.q_a_proj.input_quantizer TensorQuantizer(disabled) +model.layers.44.self_attn.q_a_proj.output_quantizer TensorQuantizer(disabled) +model.layers.44.self_attn.q_a_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.44.self_attn.q_b_proj.input_quantizer TensorQuantizer(disabled) +model.layers.44.self_attn.q_b_proj.output_quantizer TensorQuantizer(disabled) +model.layers.44.self_attn.q_b_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.44.self_attn.kv_a_proj_with_mqa.input_quantizer TensorQuantizer(disabled) +model.layers.44.self_attn.kv_a_proj_with_mqa.output_quantizer TensorQuantizer(disabled) +model.layers.44.self_attn.kv_a_proj_with_mqa.weight_quantizer TensorQuantizer(disabled) +model.layers.44.self_attn.kv_b_proj.input_quantizer TensorQuantizer(disabled) +model.layers.44.self_attn.kv_b_proj.output_quantizer TensorQuantizer(disabled) +model.layers.44.self_attn.kv_b_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.44.self_attn.o_proj.input_quantizer TensorQuantizer(disabled) +model.layers.44.self_attn.o_proj.output_quantizer TensorQuantizer(disabled) +model.layers.44.self_attn.o_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.44.self_attn.q_bmm_quantizer TensorQuantizer(disabled) +model.layers.44.self_attn.k_bmm_quantizer TensorQuantizer((4, 3) bit fake per-tensor amax=4.48e+02(const) calibrator=MaxCalibrator quant) +model.layers.44.self_attn.v_bmm_quantizer TensorQuantizer((4, 3) bit fake per-tensor amax=4.48e+02(const) calibrator=MaxCalibrator quant) +model.layers.44.self_attn.p_bmm_quantizer TensorQuantizer(disabled) +model.layers.44.mlp.experts.gate_up_proj_input_quantizer TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.05e+00 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.gate_up_proj_weight_quantizers.0 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.gate_up_proj_weight_quantizers.1 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.78e-01 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.gate_up_proj_weight_quantizers.2 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.gate_up_proj_weight_quantizers.3 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.24e-01 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.gate_up_proj_weight_quantizers.4 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.46e-01 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.gate_up_proj_weight_quantizers.5 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.20e-01 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.gate_up_proj_weight_quantizers.6 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.gate_up_proj_weight_quantizers.7 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.13e-01 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.gate_up_proj_weight_quantizers.8 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.gate_up_proj_weight_quantizers.9 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.15e-01 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.gate_up_proj_weight_quantizers.10 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=4.43e-01 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.gate_up_proj_weight_quantizers.11 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.gate_up_proj_weight_quantizers.12 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.28e-01 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.gate_up_proj_weight_quantizers.13 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.gate_up_proj_weight_quantizers.14 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.gate_up_proj_weight_quantizers.15 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.68e-01 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.gate_up_proj_weight_quantizers.16 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.81e-01 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.gate_up_proj_weight_quantizers.17 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.gate_up_proj_weight_quantizers.18 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.13e-01 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.gate_up_proj_weight_quantizers.19 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.gate_up_proj_weight_quantizers.20 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.20e-01 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.gate_up_proj_weight_quantizers.21 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.75e-01 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.gate_up_proj_weight_quantizers.22 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.gate_up_proj_weight_quantizers.23 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.53e-01 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.gate_up_proj_weight_quantizers.24 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.gate_up_proj_weight_quantizers.25 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.gate_up_proj_weight_quantizers.26 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.gate_up_proj_weight_quantizers.27 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.gate_up_proj_weight_quantizers.28 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.gate_up_proj_weight_quantizers.29 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.25e-01 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.gate_up_proj_weight_quantizers.30 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.79e-01 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.gate_up_proj_weight_quantizers.31 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.15e-01 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.gate_up_proj_weight_quantizers.32 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.gate_up_proj_weight_quantizers.33 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.24e-01 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.gate_up_proj_weight_quantizers.34 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.gate_up_proj_weight_quantizers.35 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.gate_up_proj_weight_quantizers.36 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.gate_up_proj_weight_quantizers.37 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.gate_up_proj_weight_quantizers.38 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.gate_up_proj_weight_quantizers.39 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.gate_up_proj_weight_quantizers.40 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.gate_up_proj_weight_quantizers.41 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.gate_up_proj_weight_quantizers.42 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.25e-01 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.gate_up_proj_weight_quantizers.43 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.gate_up_proj_weight_quantizers.44 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.gate_up_proj_weight_quantizers.45 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.gate_up_proj_weight_quantizers.46 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.29e-01 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.gate_up_proj_weight_quantizers.47 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.31e-01 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.gate_up_proj_weight_quantizers.48 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.13e-01 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.gate_up_proj_weight_quantizers.49 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.80e-01 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.gate_up_proj_weight_quantizers.50 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.gate_up_proj_weight_quantizers.51 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.gate_up_proj_weight_quantizers.52 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.41e-01 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.gate_up_proj_weight_quantizers.53 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.gate_up_proj_weight_quantizers.54 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.47e-01 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.gate_up_proj_weight_quantizers.55 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.57e-01 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.gate_up_proj_weight_quantizers.56 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.gate_up_proj_weight_quantizers.57 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.gate_up_proj_weight_quantizers.58 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.gate_up_proj_weight_quantizers.59 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.35e-01 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.gate_up_proj_weight_quantizers.60 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.38e-01 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.gate_up_proj_weight_quantizers.61 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.18e-01 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.gate_up_proj_weight_quantizers.62 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.gate_up_proj_weight_quantizers.63 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.gate_up_proj_weight_quantizers.64 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.14e-01 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.gate_up_proj_weight_quantizers.65 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.gate_up_proj_weight_quantizers.66 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.gate_up_proj_weight_quantizers.67 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.gate_up_proj_weight_quantizers.68 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.gate_up_proj_weight_quantizers.69 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.gate_up_proj_weight_quantizers.70 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.18e-01 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.gate_up_proj_weight_quantizers.71 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.38e-01 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.gate_up_proj_weight_quantizers.72 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.gate_up_proj_weight_quantizers.73 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.21e-01 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.gate_up_proj_weight_quantizers.74 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.gate_up_proj_weight_quantizers.75 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.87e-01 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.gate_up_proj_weight_quantizers.76 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.gate_up_proj_weight_quantizers.77 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.gate_up_proj_weight_quantizers.78 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.33e-01 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.gate_up_proj_weight_quantizers.79 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.gate_up_proj_weight_quantizers.80 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.17e-01 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.gate_up_proj_weight_quantizers.81 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.gate_up_proj_weight_quantizers.82 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.gate_up_proj_weight_quantizers.83 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.19e-01 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.gate_up_proj_weight_quantizers.84 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.64e-02 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.gate_up_proj_weight_quantizers.85 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.gate_up_proj_weight_quantizers.86 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.34e-01 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.gate_up_proj_weight_quantizers.87 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.gate_up_proj_weight_quantizers.88 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.30e-01 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.gate_up_proj_weight_quantizers.89 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.gate_up_proj_weight_quantizers.90 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.gate_up_proj_weight_quantizers.91 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.gate_up_proj_weight_quantizers.92 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.gate_up_proj_weight_quantizers.93 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.gate_up_proj_weight_quantizers.94 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.gate_up_proj_weight_quantizers.95 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.gate_up_proj_weight_quantizers.96 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.gate_up_proj_weight_quantizers.97 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.gate_up_proj_weight_quantizers.98 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.45e-01 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.gate_up_proj_weight_quantizers.99 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.gate_up_proj_weight_quantizers.100 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.23e-01 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.gate_up_proj_weight_quantizers.101 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.15e-01 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.gate_up_proj_weight_quantizers.102 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.22e-01 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.gate_up_proj_weight_quantizers.103 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.19e-01 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.gate_up_proj_weight_quantizers.104 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.21e-01 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.gate_up_proj_weight_quantizers.105 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.gate_up_proj_weight_quantizers.106 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.gate_up_proj_weight_quantizers.107 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.gate_up_proj_weight_quantizers.108 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.15e-01 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.gate_up_proj_weight_quantizers.109 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.64e-01 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.gate_up_proj_weight_quantizers.110 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.gate_up_proj_weight_quantizers.111 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.88e-01 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.gate_up_proj_weight_quantizers.112 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.19e-01 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.gate_up_proj_weight_quantizers.113 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.47e-01 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.gate_up_proj_weight_quantizers.114 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.gate_up_proj_weight_quantizers.115 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.28e-01 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.gate_up_proj_weight_quantizers.116 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.13e-01 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.gate_up_proj_weight_quantizers.117 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.gate_up_proj_weight_quantizers.118 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.52e-01 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.gate_up_proj_weight_quantizers.119 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.gate_up_proj_weight_quantizers.120 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.gate_up_proj_weight_quantizers.121 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.gate_up_proj_weight_quantizers.122 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.gate_up_proj_weight_quantizers.123 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.gate_up_proj_weight_quantizers.124 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.57e-01 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.gate_up_proj_weight_quantizers.125 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.gate_up_proj_weight_quantizers.126 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.16e-01 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.gate_up_proj_weight_quantizers.127 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.gate_up_proj_weight_quantizers.128 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.gate_up_proj_weight_quantizers.129 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.gate_up_proj_weight_quantizers.130 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.gate_up_proj_weight_quantizers.131 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.33e-01 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.gate_up_proj_weight_quantizers.132 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.16e-01 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.gate_up_proj_weight_quantizers.133 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.30e-01 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.gate_up_proj_weight_quantizers.134 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.gate_up_proj_weight_quantizers.135 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.gate_up_proj_weight_quantizers.136 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.98e-01 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.gate_up_proj_weight_quantizers.137 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.68e-01 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.gate_up_proj_weight_quantizers.138 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.51e-01 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.gate_up_proj_weight_quantizers.139 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.gate_up_proj_weight_quantizers.140 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.gate_up_proj_weight_quantizers.141 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.47e-01 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.gate_up_proj_weight_quantizers.142 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.gate_up_proj_weight_quantizers.143 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.gate_up_proj_weight_quantizers.144 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.gate_up_proj_weight_quantizers.145 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.40e-01 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.gate_up_proj_weight_quantizers.146 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.29e-01 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.gate_up_proj_weight_quantizers.147 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.gate_up_proj_weight_quantizers.148 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.00e-01 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.gate_up_proj_weight_quantizers.149 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.gate_up_proj_weight_quantizers.150 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.gate_up_proj_weight_quantizers.151 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.28e-01 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.gate_up_proj_weight_quantizers.152 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.gate_up_proj_weight_quantizers.153 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.57e-01 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.gate_up_proj_weight_quantizers.154 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.gate_up_proj_weight_quantizers.155 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.92e-01 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.gate_up_proj_weight_quantizers.156 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.87e-01 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.gate_up_proj_weight_quantizers.157 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.gate_up_proj_weight_quantizers.158 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.gate_up_proj_weight_quantizers.159 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.67e-01 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.gate_up_proj_weight_quantizers.160 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.gate_up_proj_weight_quantizers.161 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.gate_up_proj_weight_quantizers.162 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.16e-01 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.gate_up_proj_weight_quantizers.163 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=4.49e-01 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.gate_up_proj_weight_quantizers.164 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.05e-01 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.gate_up_proj_weight_quantizers.165 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.gate_up_proj_weight_quantizers.166 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.gate_up_proj_weight_quantizers.167 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.74e-01 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.gate_up_proj_weight_quantizers.168 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.21e-01 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.gate_up_proj_weight_quantizers.169 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.gate_up_proj_weight_quantizers.170 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.25e-01 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.gate_up_proj_weight_quantizers.171 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.gate_up_proj_weight_quantizers.172 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.65e-01 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.gate_up_proj_weight_quantizers.173 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.gate_up_proj_weight_quantizers.174 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.92e-01 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.gate_up_proj_weight_quantizers.175 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.gate_up_proj_weight_quantizers.176 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.gate_up_proj_weight_quantizers.177 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.gate_up_proj_weight_quantizers.178 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.32e-01 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.gate_up_proj_weight_quantizers.179 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.gate_up_proj_weight_quantizers.180 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.gate_up_proj_weight_quantizers.181 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.gate_up_proj_weight_quantizers.182 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.gate_up_proj_weight_quantizers.183 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.gate_up_proj_weight_quantizers.184 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.gate_up_proj_weight_quantizers.185 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.gate_up_proj_weight_quantizers.186 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.19e-01 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.gate_up_proj_weight_quantizers.187 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.95e-01 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.gate_up_proj_weight_quantizers.188 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.25e-01 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.gate_up_proj_weight_quantizers.189 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.13e-01 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.gate_up_proj_weight_quantizers.190 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.gate_up_proj_weight_quantizers.191 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.gate_up_proj_weight_quantizers.192 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.gate_up_proj_weight_quantizers.193 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.gate_up_proj_weight_quantizers.194 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.gate_up_proj_weight_quantizers.195 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.gate_up_proj_weight_quantizers.196 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.88e-01 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.gate_up_proj_weight_quantizers.197 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.81e-01 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.gate_up_proj_weight_quantizers.198 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.08e-01 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.gate_up_proj_weight_quantizers.199 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.gate_up_proj_weight_quantizers.200 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.gate_up_proj_weight_quantizers.201 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.gate_up_proj_weight_quantizers.202 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.30e-01 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.gate_up_proj_weight_quantizers.203 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.gate_up_proj_weight_quantizers.204 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.20e-01 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.gate_up_proj_weight_quantizers.205 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.15e-01 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.gate_up_proj_weight_quantizers.206 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.16e-01 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.gate_up_proj_weight_quantizers.207 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.gate_up_proj_weight_quantizers.208 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.gate_up_proj_weight_quantizers.209 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.gate_up_proj_weight_quantizers.210 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.gate_up_proj_weight_quantizers.211 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=4.32e-01 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.gate_up_proj_weight_quantizers.212 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.gate_up_proj_weight_quantizers.213 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.gate_up_proj_weight_quantizers.214 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.19e-01 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.gate_up_proj_weight_quantizers.215 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.gate_up_proj_weight_quantizers.216 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.gate_up_proj_weight_quantizers.217 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.50e-02 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.gate_up_proj_weight_quantizers.218 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.gate_up_proj_weight_quantizers.219 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.13e-01 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.gate_up_proj_weight_quantizers.220 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.32e-01 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.gate_up_proj_weight_quantizers.221 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.gate_up_proj_weight_quantizers.222 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.24e-01 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.gate_up_proj_weight_quantizers.223 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.32e-01 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.gate_up_proj_weight_quantizers.224 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.98e-01 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.gate_up_proj_weight_quantizers.225 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.gate_up_proj_weight_quantizers.226 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.gate_up_proj_weight_quantizers.227 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.13e-01 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.gate_up_proj_weight_quantizers.228 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.gate_up_proj_weight_quantizers.229 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.gate_up_proj_weight_quantizers.230 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.21e-01 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.gate_up_proj_weight_quantizers.231 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.22e-01 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.gate_up_proj_weight_quantizers.232 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.83e-01 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.gate_up_proj_weight_quantizers.233 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.gate_up_proj_weight_quantizers.234 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.gate_up_proj_weight_quantizers.235 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.54e-01 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.gate_up_proj_weight_quantizers.236 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.gate_up_proj_weight_quantizers.237 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.39e-01 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.gate_up_proj_weight_quantizers.238 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.97e-01 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.gate_up_proj_weight_quantizers.239 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.gate_up_proj_weight_quantizers.240 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.gate_up_proj_weight_quantizers.241 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.gate_up_proj_weight_quantizers.242 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.20e-01 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.gate_up_proj_weight_quantizers.243 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.13e-01 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.gate_up_proj_weight_quantizers.244 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.33e-01 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.gate_up_proj_weight_quantizers.245 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.gate_up_proj_weight_quantizers.246 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.gate_up_proj_weight_quantizers.247 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.gate_up_proj_weight_quantizers.248 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.gate_up_proj_weight_quantizers.249 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.gate_up_proj_weight_quantizers.250 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.gate_up_proj_weight_quantizers.251 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.16e-01 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.gate_up_proj_weight_quantizers.252 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.gate_up_proj_weight_quantizers.253 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.gate_up_proj_weight_quantizers.254 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.gate_up_proj_weight_quantizers.255 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.14e-01 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.down_proj_input_quantizer TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.98e+01 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.down_proj_weight_quantizers.0 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.down_proj_weight_quantizers.1 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.20e-01 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.down_proj_weight_quantizers.2 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.down_proj_weight_quantizers.3 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.down_proj_weight_quantizers.4 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.38e-01 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.down_proj_weight_quantizers.5 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=5.12e-01 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.down_proj_weight_quantizers.6 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.down_proj_weight_quantizers.7 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.down_proj_weight_quantizers.8 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.down_proj_weight_quantizers.9 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.down_proj_weight_quantizers.10 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.77e-01 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.down_proj_weight_quantizers.11 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.down_proj_weight_quantizers.12 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.down_proj_weight_quantizers.13 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.down_proj_weight_quantizers.14 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.60e-01 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.down_proj_weight_quantizers.15 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.43e-01 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.down_proj_weight_quantizers.16 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.down_proj_weight_quantizers.17 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.down_proj_weight_quantizers.18 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.down_proj_weight_quantizers.19 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.down_proj_weight_quantizers.20 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.70e-01 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.down_proj_weight_quantizers.21 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.down_proj_weight_quantizers.22 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.down_proj_weight_quantizers.23 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.37e-01 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.down_proj_weight_quantizers.24 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.down_proj_weight_quantizers.25 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.down_proj_weight_quantizers.26 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.down_proj_weight_quantizers.27 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.down_proj_weight_quantizers.28 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.64e-02 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.down_proj_weight_quantizers.29 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.64e-02 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.down_proj_weight_quantizers.30 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.down_proj_weight_quantizers.31 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.down_proj_weight_quantizers.32 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.down_proj_weight_quantizers.33 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.71e-01 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.down_proj_weight_quantizers.34 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.down_proj_weight_quantizers.35 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.down_proj_weight_quantizers.36 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.17e-01 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.down_proj_weight_quantizers.37 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.down_proj_weight_quantizers.38 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.down_proj_weight_quantizers.39 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.down_proj_weight_quantizers.40 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.11e-01 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.down_proj_weight_quantizers.41 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.40e-02 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.down_proj_weight_quantizers.42 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.down_proj_weight_quantizers.43 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.down_proj_weight_quantizers.44 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.down_proj_weight_quantizers.45 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.down_proj_weight_quantizers.46 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.down_proj_weight_quantizers.47 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.down_proj_weight_quantizers.48 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.04e-01 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.down_proj_weight_quantizers.49 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.38e-01 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.down_proj_weight_quantizers.50 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.down_proj_weight_quantizers.51 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.down_proj_weight_quantizers.52 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.35e-02 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.down_proj_weight_quantizers.53 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.11e-01 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.down_proj_weight_quantizers.54 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.down_proj_weight_quantizers.55 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.83e-01 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.down_proj_weight_quantizers.56 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.down_proj_weight_quantizers.57 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.down_proj_weight_quantizers.58 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.64e-02 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.down_proj_weight_quantizers.59 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.down_proj_weight_quantizers.60 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.down_proj_weight_quantizers.61 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.down_proj_weight_quantizers.62 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.64e-01 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.down_proj_weight_quantizers.63 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.down_proj_weight_quantizers.64 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.59e-02 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.down_proj_weight_quantizers.65 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.down_proj_weight_quantizers.66 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.down_proj_weight_quantizers.67 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.down_proj_weight_quantizers.68 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.14e-01 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.down_proj_weight_quantizers.69 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.down_proj_weight_quantizers.70 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.down_proj_weight_quantizers.71 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.28e-01 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.down_proj_weight_quantizers.72 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.down_proj_weight_quantizers.73 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.down_proj_weight_quantizers.74 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.45e-02 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.down_proj_weight_quantizers.75 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.down_proj_weight_quantizers.76 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.down_proj_weight_quantizers.77 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.down_proj_weight_quantizers.78 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.down_proj_weight_quantizers.79 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.down_proj_weight_quantizers.80 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.down_proj_weight_quantizers.81 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.down_proj_weight_quantizers.82 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.down_proj_weight_quantizers.83 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.50e-02 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.down_proj_weight_quantizers.84 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.64e-02 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.down_proj_weight_quantizers.85 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.down_proj_weight_quantizers.86 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.down_proj_weight_quantizers.87 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.29e-01 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.down_proj_weight_quantizers.88 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.down_proj_weight_quantizers.89 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.down_proj_weight_quantizers.90 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.85e-01 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.down_proj_weight_quantizers.91 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.down_proj_weight_quantizers.92 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.down_proj_weight_quantizers.93 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.54e-02 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.down_proj_weight_quantizers.94 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.64e-02 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.down_proj_weight_quantizers.95 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.down_proj_weight_quantizers.96 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.down_proj_weight_quantizers.97 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.down_proj_weight_quantizers.98 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.down_proj_weight_quantizers.99 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.down_proj_weight_quantizers.100 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.down_proj_weight_quantizers.101 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.down_proj_weight_quantizers.102 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.down_proj_weight_quantizers.103 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.down_proj_weight_quantizers.104 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.down_proj_weight_quantizers.105 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.down_proj_weight_quantizers.106 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.down_proj_weight_quantizers.107 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.22e-01 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.down_proj_weight_quantizers.108 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.down_proj_weight_quantizers.109 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.down_proj_weight_quantizers.110 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.down_proj_weight_quantizers.111 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.86e-01 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.down_proj_weight_quantizers.112 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.down_proj_weight_quantizers.113 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.94e-01 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.down_proj_weight_quantizers.114 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.down_proj_weight_quantizers.115 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.41e-01 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.down_proj_weight_quantizers.116 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.29e-01 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.down_proj_weight_quantizers.117 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.59e-02 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.down_proj_weight_quantizers.118 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.13e-01 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.down_proj_weight_quantizers.119 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.62e-01 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.down_proj_weight_quantizers.120 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.54e-02 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.down_proj_weight_quantizers.121 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.down_proj_weight_quantizers.122 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.down_proj_weight_quantizers.123 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.down_proj_weight_quantizers.124 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.81e-01 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.down_proj_weight_quantizers.125 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.down_proj_weight_quantizers.126 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.down_proj_weight_quantizers.127 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.59e-02 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.down_proj_weight_quantizers.128 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.down_proj_weight_quantizers.129 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.down_proj_weight_quantizers.130 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.down_proj_weight_quantizers.131 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.01e-01 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.down_proj_weight_quantizers.132 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.down_proj_weight_quantizers.133 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.down_proj_weight_quantizers.134 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.down_proj_weight_quantizers.135 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.down_proj_weight_quantizers.136 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.down_proj_weight_quantizers.137 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.72e-01 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.down_proj_weight_quantizers.138 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.62e-01 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.down_proj_weight_quantizers.139 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.down_proj_weight_quantizers.140 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.down_proj_weight_quantizers.141 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.down_proj_weight_quantizers.142 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.58e-01 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.down_proj_weight_quantizers.143 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.down_proj_weight_quantizers.144 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.down_proj_weight_quantizers.145 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.down_proj_weight_quantizers.146 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.03e-01 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.down_proj_weight_quantizers.147 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.down_proj_weight_quantizers.148 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.down_proj_weight_quantizers.149 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.down_proj_weight_quantizers.150 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.down_proj_weight_quantizers.151 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.72e-01 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.down_proj_weight_quantizers.152 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.down_proj_weight_quantizers.153 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.54e-02 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.down_proj_weight_quantizers.154 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.50e-02 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.down_proj_weight_quantizers.155 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.21e-01 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.down_proj_weight_quantizers.156 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.down_proj_weight_quantizers.157 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.21e-01 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.down_proj_weight_quantizers.158 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.down_proj_weight_quantizers.159 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.down_proj_weight_quantizers.160 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.down_proj_weight_quantizers.161 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.down_proj_weight_quantizers.162 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.down_proj_weight_quantizers.163 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.52e-01 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.down_proj_weight_quantizers.164 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.42e-01 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.down_proj_weight_quantizers.165 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.down_proj_weight_quantizers.166 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.down_proj_weight_quantizers.167 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.59e-02 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.down_proj_weight_quantizers.168 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.down_proj_weight_quantizers.169 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.50e-02 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.down_proj_weight_quantizers.170 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.down_proj_weight_quantizers.171 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.down_proj_weight_quantizers.172 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.down_proj_weight_quantizers.173 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.down_proj_weight_quantizers.174 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.down_proj_weight_quantizers.175 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.down_proj_weight_quantizers.176 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.down_proj_weight_quantizers.177 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.59e-02 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.down_proj_weight_quantizers.178 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.down_proj_weight_quantizers.179 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.down_proj_weight_quantizers.180 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.35e-02 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.down_proj_weight_quantizers.181 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.down_proj_weight_quantizers.182 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.64e-02 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.down_proj_weight_quantizers.183 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.down_proj_weight_quantizers.184 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.down_proj_weight_quantizers.185 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.down_proj_weight_quantizers.186 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.down_proj_weight_quantizers.187 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.03e-01 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.down_proj_weight_quantizers.188 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.17e-01 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.down_proj_weight_quantizers.189 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.down_proj_weight_quantizers.190 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.19e-01 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.down_proj_weight_quantizers.191 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.down_proj_weight_quantizers.192 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.down_proj_weight_quantizers.193 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.down_proj_weight_quantizers.194 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.down_proj_weight_quantizers.195 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.down_proj_weight_quantizers.196 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.24e-01 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.down_proj_weight_quantizers.197 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.20e-01 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.down_proj_weight_quantizers.198 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.down_proj_weight_quantizers.199 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.59e-02 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.down_proj_weight_quantizers.200 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.down_proj_weight_quantizers.201 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.down_proj_weight_quantizers.202 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.67e-01 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.down_proj_weight_quantizers.203 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.down_proj_weight_quantizers.204 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.66e-01 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.down_proj_weight_quantizers.205 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.down_proj_weight_quantizers.206 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.down_proj_weight_quantizers.207 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.50e-02 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.down_proj_weight_quantizers.208 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.down_proj_weight_quantizers.209 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.45e-02 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.down_proj_weight_quantizers.210 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.29e-01 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.down_proj_weight_quantizers.211 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.35e-01 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.down_proj_weight_quantizers.212 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.32e-01 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.down_proj_weight_quantizers.213 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.down_proj_weight_quantizers.214 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.down_proj_weight_quantizers.215 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.down_proj_weight_quantizers.216 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.down_proj_weight_quantizers.217 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.down_proj_weight_quantizers.218 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.down_proj_weight_quantizers.219 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.down_proj_weight_quantizers.220 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.down_proj_weight_quantizers.221 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.down_proj_weight_quantizers.222 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.54e-02 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.down_proj_weight_quantizers.223 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.18e-01 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.down_proj_weight_quantizers.224 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.13e-01 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.down_proj_weight_quantizers.225 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.down_proj_weight_quantizers.226 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.down_proj_weight_quantizers.227 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.46e-01 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.down_proj_weight_quantizers.228 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.down_proj_weight_quantizers.229 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.down_proj_weight_quantizers.230 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.18e-01 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.down_proj_weight_quantizers.231 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.down_proj_weight_quantizers.232 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.75e-01 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.down_proj_weight_quantizers.233 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.down_proj_weight_quantizers.234 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.down_proj_weight_quantizers.235 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.01e-01 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.down_proj_weight_quantizers.236 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.down_proj_weight_quantizers.237 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.27e-01 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.down_proj_weight_quantizers.238 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.40e-02 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.down_proj_weight_quantizers.239 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.down_proj_weight_quantizers.240 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.down_proj_weight_quantizers.241 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.29e-01 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.down_proj_weight_quantizers.242 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.down_proj_weight_quantizers.243 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.down_proj_weight_quantizers.244 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.down_proj_weight_quantizers.245 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.09e-01 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.down_proj_weight_quantizers.246 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.down_proj_weight_quantizers.247 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.down_proj_weight_quantizers.248 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.54e-02 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.down_proj_weight_quantizers.249 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.59e-02 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.down_proj_weight_quantizers.250 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.down_proj_weight_quantizers.251 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.down_proj_weight_quantizers.252 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.down_proj_weight_quantizers.253 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.down_proj_weight_quantizers.254 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.down_proj_weight_quantizers.255 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.44.mlp.shared_experts.gate_proj.input_quantizer TensorQuantizer(disabled) +model.layers.44.mlp.shared_experts.gate_proj.output_quantizer TensorQuantizer(disabled) +model.layers.44.mlp.shared_experts.gate_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.44.mlp.shared_experts.up_proj.input_quantizer TensorQuantizer(disabled) +model.layers.44.mlp.shared_experts.up_proj.output_quantizer TensorQuantizer(disabled) +model.layers.44.mlp.shared_experts.up_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.44.mlp.shared_experts.down_proj.input_quantizer TensorQuantizer(disabled) +model.layers.44.mlp.shared_experts.down_proj.output_quantizer TensorQuantizer(disabled) +model.layers.44.mlp.shared_experts.down_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.45.self_attn.q_a_proj.input_quantizer TensorQuantizer(disabled) +model.layers.45.self_attn.q_a_proj.output_quantizer TensorQuantizer(disabled) +model.layers.45.self_attn.q_a_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.45.self_attn.q_b_proj.input_quantizer TensorQuantizer(disabled) +model.layers.45.self_attn.q_b_proj.output_quantizer TensorQuantizer(disabled) +model.layers.45.self_attn.q_b_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.45.self_attn.kv_a_proj_with_mqa.input_quantizer TensorQuantizer(disabled) +model.layers.45.self_attn.kv_a_proj_with_mqa.output_quantizer TensorQuantizer(disabled) +model.layers.45.self_attn.kv_a_proj_with_mqa.weight_quantizer TensorQuantizer(disabled) +model.layers.45.self_attn.kv_b_proj.input_quantizer TensorQuantizer(disabled) +model.layers.45.self_attn.kv_b_proj.output_quantizer TensorQuantizer(disabled) +model.layers.45.self_attn.kv_b_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.45.self_attn.o_proj.input_quantizer TensorQuantizer(disabled) +model.layers.45.self_attn.o_proj.output_quantizer TensorQuantizer(disabled) +model.layers.45.self_attn.o_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.45.self_attn.q_bmm_quantizer TensorQuantizer(disabled) +model.layers.45.self_attn.k_bmm_quantizer TensorQuantizer((4, 3) bit fake per-tensor amax=4.48e+02(const) calibrator=MaxCalibrator quant) +model.layers.45.self_attn.v_bmm_quantizer TensorQuantizer((4, 3) bit fake per-tensor amax=4.48e+02(const) calibrator=MaxCalibrator quant) +model.layers.45.self_attn.p_bmm_quantizer TensorQuantizer(disabled) +model.layers.45.mlp.experts.gate_up_proj_input_quantizer TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.23e+00 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.gate_up_proj_weight_quantizers.0 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.37e-01 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.gate_up_proj_weight_quantizers.1 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.17e-01 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.gate_up_proj_weight_quantizers.2 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.gate_up_proj_weight_quantizers.3 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.gate_up_proj_weight_quantizers.4 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.29e-01 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.gate_up_proj_weight_quantizers.5 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.gate_up_proj_weight_quantizers.6 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.gate_up_proj_weight_quantizers.7 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.21e-01 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.gate_up_proj_weight_quantizers.8 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.gate_up_proj_weight_quantizers.9 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.22e-01 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.gate_up_proj_weight_quantizers.10 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.27e-01 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.gate_up_proj_weight_quantizers.11 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.gate_up_proj_weight_quantizers.12 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.gate_up_proj_weight_quantizers.13 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.gate_up_proj_weight_quantizers.14 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.24e-01 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.gate_up_proj_weight_quantizers.15 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.34e-01 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.gate_up_proj_weight_quantizers.16 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.15e-01 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.gate_up_proj_weight_quantizers.17 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.20e-01 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.gate_up_proj_weight_quantizers.18 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.52e-01 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.gate_up_proj_weight_quantizers.19 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.62e-01 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.gate_up_proj_weight_quantizers.20 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.gate_up_proj_weight_quantizers.21 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.gate_up_proj_weight_quantizers.22 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.59e-01 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.gate_up_proj_weight_quantizers.23 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.gate_up_proj_weight_quantizers.24 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.33e-01 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.gate_up_proj_weight_quantizers.25 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.gate_up_proj_weight_quantizers.26 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.82e-01 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.gate_up_proj_weight_quantizers.27 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.78e-01 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.gate_up_proj_weight_quantizers.28 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.gate_up_proj_weight_quantizers.29 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.27e-01 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.gate_up_proj_weight_quantizers.30 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.gate_up_proj_weight_quantizers.31 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.gate_up_proj_weight_quantizers.32 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.gate_up_proj_weight_quantizers.33 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.13e-01 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.gate_up_proj_weight_quantizers.34 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.22e-01 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.gate_up_proj_weight_quantizers.35 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.81e-01 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.gate_up_proj_weight_quantizers.36 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.48e-01 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.gate_up_proj_weight_quantizers.37 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.gate_up_proj_weight_quantizers.38 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.17e-01 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.gate_up_proj_weight_quantizers.39 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.gate_up_proj_weight_quantizers.40 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.gate_up_proj_weight_quantizers.41 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.21e-01 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.gate_up_proj_weight_quantizers.42 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.gate_up_proj_weight_quantizers.43 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.gate_up_proj_weight_quantizers.44 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.gate_up_proj_weight_quantizers.45 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.52e-01 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.gate_up_proj_weight_quantizers.46 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.gate_up_proj_weight_quantizers.47 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.gate_up_proj_weight_quantizers.48 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.14e-01 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.gate_up_proj_weight_quantizers.49 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.18e-01 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.gate_up_proj_weight_quantizers.50 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.gate_up_proj_weight_quantizers.51 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.23e-01 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.gate_up_proj_weight_quantizers.52 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.gate_up_proj_weight_quantizers.53 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.gate_up_proj_weight_quantizers.54 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.36e-01 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.gate_up_proj_weight_quantizers.55 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.gate_up_proj_weight_quantizers.56 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.39e-01 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.gate_up_proj_weight_quantizers.57 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.29e-01 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.gate_up_proj_weight_quantizers.58 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.14e-01 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.gate_up_proj_weight_quantizers.59 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.45e-01 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.gate_up_proj_weight_quantizers.60 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.50e-01 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.gate_up_proj_weight_quantizers.61 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.31e-01 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.gate_up_proj_weight_quantizers.62 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.gate_up_proj_weight_quantizers.63 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.gate_up_proj_weight_quantizers.64 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.gate_up_proj_weight_quantizers.65 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.22e-01 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.gate_up_proj_weight_quantizers.66 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.gate_up_proj_weight_quantizers.67 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.gate_up_proj_weight_quantizers.68 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.gate_up_proj_weight_quantizers.69 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.54e-01 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.gate_up_proj_weight_quantizers.70 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.gate_up_proj_weight_quantizers.71 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.22e-01 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.gate_up_proj_weight_quantizers.72 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.gate_up_proj_weight_quantizers.73 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.gate_up_proj_weight_quantizers.74 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.55e-01 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.gate_up_proj_weight_quantizers.75 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.gate_up_proj_weight_quantizers.76 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.gate_up_proj_weight_quantizers.77 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.31e-01 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.gate_up_proj_weight_quantizers.78 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.gate_up_proj_weight_quantizers.79 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.37e-01 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.gate_up_proj_weight_quantizers.80 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.16e-01 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.gate_up_proj_weight_quantizers.81 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.81e-01 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.gate_up_proj_weight_quantizers.82 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.gate_up_proj_weight_quantizers.83 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.gate_up_proj_weight_quantizers.84 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.gate_up_proj_weight_quantizers.85 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.gate_up_proj_weight_quantizers.86 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.gate_up_proj_weight_quantizers.87 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.14e-01 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.gate_up_proj_weight_quantizers.88 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.gate_up_proj_weight_quantizers.89 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.13e-01 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.gate_up_proj_weight_quantizers.90 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.84e-01 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.gate_up_proj_weight_quantizers.91 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.gate_up_proj_weight_quantizers.92 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.gate_up_proj_weight_quantizers.93 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.54e-02 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.gate_up_proj_weight_quantizers.94 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.gate_up_proj_weight_quantizers.95 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.gate_up_proj_weight_quantizers.96 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.gate_up_proj_weight_quantizers.97 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.gate_up_proj_weight_quantizers.98 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.gate_up_proj_weight_quantizers.99 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.gate_up_proj_weight_quantizers.100 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.23e-01 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.gate_up_proj_weight_quantizers.101 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.gate_up_proj_weight_quantizers.102 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.46e-01 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.gate_up_proj_weight_quantizers.103 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.20e-01 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.gate_up_proj_weight_quantizers.104 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.34e-01 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.gate_up_proj_weight_quantizers.105 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.24e-01 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.gate_up_proj_weight_quantizers.106 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.58e-01 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.gate_up_proj_weight_quantizers.107 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.21e-01 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.gate_up_proj_weight_quantizers.108 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.gate_up_proj_weight_quantizers.109 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.20e-01 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.gate_up_proj_weight_quantizers.110 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.22e-01 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.gate_up_proj_weight_quantizers.111 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.22e-01 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.gate_up_proj_weight_quantizers.112 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.07e-01 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.gate_up_proj_weight_quantizers.113 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.13e-01 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.gate_up_proj_weight_quantizers.114 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.27e-01 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.gate_up_proj_weight_quantizers.115 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.15e-01 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.gate_up_proj_weight_quantizers.116 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.34e-01 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.gate_up_proj_weight_quantizers.117 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.gate_up_proj_weight_quantizers.118 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.57e-01 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.gate_up_proj_weight_quantizers.119 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.gate_up_proj_weight_quantizers.120 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.gate_up_proj_weight_quantizers.121 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.31e-01 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.gate_up_proj_weight_quantizers.122 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.28e-01 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.gate_up_proj_weight_quantizers.123 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.85e-01 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.gate_up_proj_weight_quantizers.124 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.33e-01 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.gate_up_proj_weight_quantizers.125 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.gate_up_proj_weight_quantizers.126 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.41e-01 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.gate_up_proj_weight_quantizers.127 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.gate_up_proj_weight_quantizers.128 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.gate_up_proj_weight_quantizers.129 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.gate_up_proj_weight_quantizers.130 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.24e-01 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.gate_up_proj_weight_quantizers.131 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.gate_up_proj_weight_quantizers.132 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.22e-01 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.gate_up_proj_weight_quantizers.133 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.gate_up_proj_weight_quantizers.134 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.26e-01 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.gate_up_proj_weight_quantizers.135 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.15e-01 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.gate_up_proj_weight_quantizers.136 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.gate_up_proj_weight_quantizers.137 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.69e-01 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.gate_up_proj_weight_quantizers.138 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.gate_up_proj_weight_quantizers.139 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.25e-01 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.gate_up_proj_weight_quantizers.140 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.gate_up_proj_weight_quantizers.141 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=4.30e-01 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.gate_up_proj_weight_quantizers.142 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.25e-01 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.gate_up_proj_weight_quantizers.143 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.16e-01 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.gate_up_proj_weight_quantizers.144 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.27e-01 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.gate_up_proj_weight_quantizers.145 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.gate_up_proj_weight_quantizers.146 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.46e-01 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.gate_up_proj_weight_quantizers.147 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.gate_up_proj_weight_quantizers.148 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.gate_up_proj_weight_quantizers.149 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.19e-01 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.gate_up_proj_weight_quantizers.150 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.77e-01 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.gate_up_proj_weight_quantizers.151 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.gate_up_proj_weight_quantizers.152 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.gate_up_proj_weight_quantizers.153 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.gate_up_proj_weight_quantizers.154 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.gate_up_proj_weight_quantizers.155 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.33e-01 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.gate_up_proj_weight_quantizers.156 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.gate_up_proj_weight_quantizers.157 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.gate_up_proj_weight_quantizers.158 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.37e-01 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.gate_up_proj_weight_quantizers.159 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.gate_up_proj_weight_quantizers.160 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.gate_up_proj_weight_quantizers.161 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.01e-01 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.gate_up_proj_weight_quantizers.162 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.gate_up_proj_weight_quantizers.163 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.gate_up_proj_weight_quantizers.164 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.gate_up_proj_weight_quantizers.165 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.gate_up_proj_weight_quantizers.166 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.54e-01 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.gate_up_proj_weight_quantizers.167 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=5.04e-01 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.gate_up_proj_weight_quantizers.168 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.68e-01 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.gate_up_proj_weight_quantizers.169 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.37e-01 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.gate_up_proj_weight_quantizers.170 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.gate_up_proj_weight_quantizers.171 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.gate_up_proj_weight_quantizers.172 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.31e-01 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.gate_up_proj_weight_quantizers.173 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.gate_up_proj_weight_quantizers.174 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.35e-01 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.gate_up_proj_weight_quantizers.175 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.15e-01 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.gate_up_proj_weight_quantizers.176 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.gate_up_proj_weight_quantizers.177 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.gate_up_proj_weight_quantizers.178 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.gate_up_proj_weight_quantizers.179 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.gate_up_proj_weight_quantizers.180 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.18e-01 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.gate_up_proj_weight_quantizers.181 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.75e-01 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.gate_up_proj_weight_quantizers.182 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.60e-01 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.gate_up_proj_weight_quantizers.183 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.gate_up_proj_weight_quantizers.184 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.gate_up_proj_weight_quantizers.185 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.25e-01 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.gate_up_proj_weight_quantizers.186 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.gate_up_proj_weight_quantizers.187 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.32e-01 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.gate_up_proj_weight_quantizers.188 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.25e-01 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.gate_up_proj_weight_quantizers.189 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.gate_up_proj_weight_quantizers.190 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.27e-01 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.gate_up_proj_weight_quantizers.191 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.gate_up_proj_weight_quantizers.192 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.29e-01 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.gate_up_proj_weight_quantizers.193 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.22e-01 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.gate_up_proj_weight_quantizers.194 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.17e-01 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.gate_up_proj_weight_quantizers.195 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.gate_up_proj_weight_quantizers.196 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.gate_up_proj_weight_quantizers.197 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.gate_up_proj_weight_quantizers.198 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.gate_up_proj_weight_quantizers.199 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.gate_up_proj_weight_quantizers.200 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.86e-01 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.gate_up_proj_weight_quantizers.201 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.20e-01 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.gate_up_proj_weight_quantizers.202 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.gate_up_proj_weight_quantizers.203 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.gate_up_proj_weight_quantizers.204 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.gate_up_proj_weight_quantizers.205 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.gate_up_proj_weight_quantizers.206 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.gate_up_proj_weight_quantizers.207 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.31e-01 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.gate_up_proj_weight_quantizers.208 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.gate_up_proj_weight_quantizers.209 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.gate_up_proj_weight_quantizers.210 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.33e-01 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.gate_up_proj_weight_quantizers.211 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.gate_up_proj_weight_quantizers.212 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.17e-01 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.gate_up_proj_weight_quantizers.213 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.gate_up_proj_weight_quantizers.214 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.17e-01 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.gate_up_proj_weight_quantizers.215 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.gate_up_proj_weight_quantizers.216 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.41e-01 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.gate_up_proj_weight_quantizers.217 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.46e-01 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.gate_up_proj_weight_quantizers.218 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.gate_up_proj_weight_quantizers.219 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.gate_up_proj_weight_quantizers.220 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.27e-01 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.gate_up_proj_weight_quantizers.221 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.gate_up_proj_weight_quantizers.222 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.gate_up_proj_weight_quantizers.223 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.gate_up_proj_weight_quantizers.224 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.gate_up_proj_weight_quantizers.225 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.25e-01 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.gate_up_proj_weight_quantizers.226 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.gate_up_proj_weight_quantizers.227 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.33e-01 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.gate_up_proj_weight_quantizers.228 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.gate_up_proj_weight_quantizers.229 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.gate_up_proj_weight_quantizers.230 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.gate_up_proj_weight_quantizers.231 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=4.20e-01 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.gate_up_proj_weight_quantizers.232 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.30e-01 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.gate_up_proj_weight_quantizers.233 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.22e-01 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.gate_up_proj_weight_quantizers.234 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.32e-01 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.gate_up_proj_weight_quantizers.235 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.34e-01 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.gate_up_proj_weight_quantizers.236 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.90e-01 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.gate_up_proj_weight_quantizers.237 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.gate_up_proj_weight_quantizers.238 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.79e-01 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.gate_up_proj_weight_quantizers.239 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.gate_up_proj_weight_quantizers.240 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.16e-01 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.gate_up_proj_weight_quantizers.241 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.15e-01 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.gate_up_proj_weight_quantizers.242 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.gate_up_proj_weight_quantizers.243 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.64e-01 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.gate_up_proj_weight_quantizers.244 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.gate_up_proj_weight_quantizers.245 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.gate_up_proj_weight_quantizers.246 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.17e-01 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.gate_up_proj_weight_quantizers.247 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.15e-01 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.gate_up_proj_weight_quantizers.248 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.23e-01 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.gate_up_proj_weight_quantizers.249 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.gate_up_proj_weight_quantizers.250 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.29e-01 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.gate_up_proj_weight_quantizers.251 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.gate_up_proj_weight_quantizers.252 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.73e-01 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.gate_up_proj_weight_quantizers.253 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.52e-01 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.gate_up_proj_weight_quantizers.254 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.gate_up_proj_weight_quantizers.255 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.down_proj_input_quantizer TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.78e+01 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.down_proj_weight_quantizers.0 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.45e-02 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.down_proj_weight_quantizers.1 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.down_proj_weight_quantizers.2 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.down_proj_weight_quantizers.3 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.down_proj_weight_quantizers.4 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.down_proj_weight_quantizers.5 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.down_proj_weight_quantizers.6 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.down_proj_weight_quantizers.7 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.down_proj_weight_quantizers.8 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.down_proj_weight_quantizers.9 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.down_proj_weight_quantizers.10 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.down_proj_weight_quantizers.11 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.45e-02 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.down_proj_weight_quantizers.12 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.down_proj_weight_quantizers.13 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.down_proj_weight_quantizers.14 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.38e-01 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.down_proj_weight_quantizers.15 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.down_proj_weight_quantizers.16 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.down_proj_weight_quantizers.17 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.13e-01 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.down_proj_weight_quantizers.18 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.down_proj_weight_quantizers.19 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.down_proj_weight_quantizers.20 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.54e-02 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.down_proj_weight_quantizers.21 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.down_proj_weight_quantizers.22 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.down_proj_weight_quantizers.23 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.37e-01 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.down_proj_weight_quantizers.24 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.down_proj_weight_quantizers.25 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.down_proj_weight_quantizers.26 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.down_proj_weight_quantizers.27 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.21e-01 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.down_proj_weight_quantizers.28 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.down_proj_weight_quantizers.29 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.22e-01 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.down_proj_weight_quantizers.30 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.64e-02 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.down_proj_weight_quantizers.31 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.down_proj_weight_quantizers.32 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.down_proj_weight_quantizers.33 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.down_proj_weight_quantizers.34 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.down_proj_weight_quantizers.35 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.19e-01 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.down_proj_weight_quantizers.36 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.48e-01 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.down_proj_weight_quantizers.37 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.down_proj_weight_quantizers.38 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.down_proj_weight_quantizers.39 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.down_proj_weight_quantizers.40 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.down_proj_weight_quantizers.41 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.down_proj_weight_quantizers.42 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.down_proj_weight_quantizers.43 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.down_proj_weight_quantizers.44 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.down_proj_weight_quantizers.45 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.down_proj_weight_quantizers.46 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.down_proj_weight_quantizers.47 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.down_proj_weight_quantizers.48 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.down_proj_weight_quantizers.49 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.down_proj_weight_quantizers.50 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.54e-02 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.down_proj_weight_quantizers.51 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.47e-01 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.down_proj_weight_quantizers.52 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.down_proj_weight_quantizers.53 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.16e-01 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.down_proj_weight_quantizers.54 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.37e-01 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.down_proj_weight_quantizers.55 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.down_proj_weight_quantizers.56 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.59e-02 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.down_proj_weight_quantizers.57 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.down_proj_weight_quantizers.58 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.down_proj_weight_quantizers.59 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.19e-01 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.down_proj_weight_quantizers.60 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.82e-01 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.down_proj_weight_quantizers.61 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.down_proj_weight_quantizers.62 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.down_proj_weight_quantizers.63 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.down_proj_weight_quantizers.64 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.down_proj_weight_quantizers.65 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.down_proj_weight_quantizers.66 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.down_proj_weight_quantizers.67 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.down_proj_weight_quantizers.68 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.down_proj_weight_quantizers.69 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.down_proj_weight_quantizers.70 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.down_proj_weight_quantizers.71 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.down_proj_weight_quantizers.72 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.down_proj_weight_quantizers.73 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.down_proj_weight_quantizers.74 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.down_proj_weight_quantizers.75 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.down_proj_weight_quantizers.76 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.18e-01 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.down_proj_weight_quantizers.77 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.down_proj_weight_quantizers.78 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.down_proj_weight_quantizers.79 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.down_proj_weight_quantizers.80 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.17e-01 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.down_proj_weight_quantizers.81 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.97e-01 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.down_proj_weight_quantizers.82 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.down_proj_weight_quantizers.83 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.down_proj_weight_quantizers.84 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.down_proj_weight_quantizers.85 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.45e-02 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.down_proj_weight_quantizers.86 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.down_proj_weight_quantizers.87 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.down_proj_weight_quantizers.88 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.down_proj_weight_quantizers.89 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.down_proj_weight_quantizers.90 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.38e-01 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.down_proj_weight_quantizers.91 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.down_proj_weight_quantizers.92 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.down_proj_weight_quantizers.93 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.down_proj_weight_quantizers.94 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.down_proj_weight_quantizers.95 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.down_proj_weight_quantizers.96 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.down_proj_weight_quantizers.97 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.down_proj_weight_quantizers.98 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.down_proj_weight_quantizers.99 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.down_proj_weight_quantizers.100 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.down_proj_weight_quantizers.101 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.down_proj_weight_quantizers.102 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.down_proj_weight_quantizers.103 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.down_proj_weight_quantizers.104 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.15e-02 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.down_proj_weight_quantizers.105 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.99e-01 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.down_proj_weight_quantizers.106 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.down_proj_weight_quantizers.107 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.59e-02 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.down_proj_weight_quantizers.108 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.64e-02 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.down_proj_weight_quantizers.109 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.16e-01 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.down_proj_weight_quantizers.110 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.down_proj_weight_quantizers.111 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.down_proj_weight_quantizers.112 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.down_proj_weight_quantizers.113 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.down_proj_weight_quantizers.114 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.down_proj_weight_quantizers.115 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.down_proj_weight_quantizers.116 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.24e-01 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.down_proj_weight_quantizers.117 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.down_proj_weight_quantizers.118 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.31e-01 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.down_proj_weight_quantizers.119 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.down_proj_weight_quantizers.120 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.down_proj_weight_quantizers.121 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.88e-01 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.down_proj_weight_quantizers.122 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.down_proj_weight_quantizers.123 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.69e-01 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.down_proj_weight_quantizers.124 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.down_proj_weight_quantizers.125 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.down_proj_weight_quantizers.126 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.down_proj_weight_quantizers.127 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.down_proj_weight_quantizers.128 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.down_proj_weight_quantizers.129 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.down_proj_weight_quantizers.130 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.27e-01 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.down_proj_weight_quantizers.131 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.down_proj_weight_quantizers.132 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.down_proj_weight_quantizers.133 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.21e-01 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.down_proj_weight_quantizers.134 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.50e-02 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.down_proj_weight_quantizers.135 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.down_proj_weight_quantizers.136 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.down_proj_weight_quantizers.137 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.56e-01 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.down_proj_weight_quantizers.138 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.down_proj_weight_quantizers.139 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=4.47e-01 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.down_proj_weight_quantizers.140 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.down_proj_weight_quantizers.141 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.29e-01 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.down_proj_weight_quantizers.142 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.down_proj_weight_quantizers.143 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.down_proj_weight_quantizers.144 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.26e-01 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.down_proj_weight_quantizers.145 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.down_proj_weight_quantizers.146 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.94e-01 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.down_proj_weight_quantizers.147 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.down_proj_weight_quantizers.148 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.down_proj_weight_quantizers.149 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.down_proj_weight_quantizers.150 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.down_proj_weight_quantizers.151 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.26e-01 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.down_proj_weight_quantizers.152 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.down_proj_weight_quantizers.153 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.down_proj_weight_quantizers.154 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.down_proj_weight_quantizers.155 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.09e-01 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.down_proj_weight_quantizers.156 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.45e-02 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.down_proj_weight_quantizers.157 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.64e-02 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.down_proj_weight_quantizers.158 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.64e-02 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.down_proj_weight_quantizers.159 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.20e-01 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.down_proj_weight_quantizers.160 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.down_proj_weight_quantizers.161 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.21e-01 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.down_proj_weight_quantizers.162 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.down_proj_weight_quantizers.163 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.down_proj_weight_quantizers.164 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.down_proj_weight_quantizers.165 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.down_proj_weight_quantizers.166 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.29e-01 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.down_proj_weight_quantizers.167 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.29e-01 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.down_proj_weight_quantizers.168 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.down_proj_weight_quantizers.169 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.50e-02 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.down_proj_weight_quantizers.170 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.down_proj_weight_quantizers.171 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.down_proj_weight_quantizers.172 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.down_proj_weight_quantizers.173 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.down_proj_weight_quantizers.174 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.down_proj_weight_quantizers.175 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.down_proj_weight_quantizers.176 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.down_proj_weight_quantizers.177 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.54e-02 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.down_proj_weight_quantizers.178 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.down_proj_weight_quantizers.179 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.down_proj_weight_quantizers.180 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.down_proj_weight_quantizers.181 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.down_proj_weight_quantizers.182 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.down_proj_weight_quantizers.183 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.down_proj_weight_quantizers.184 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.down_proj_weight_quantizers.185 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.54e-02 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.down_proj_weight_quantizers.186 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.50e-02 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.down_proj_weight_quantizers.187 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.35e-02 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.down_proj_weight_quantizers.188 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.down_proj_weight_quantizers.189 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.down_proj_weight_quantizers.190 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.down_proj_weight_quantizers.191 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.down_proj_weight_quantizers.192 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.down_proj_weight_quantizers.193 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.down_proj_weight_quantizers.194 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.down_proj_weight_quantizers.195 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.down_proj_weight_quantizers.196 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.down_proj_weight_quantizers.197 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.down_proj_weight_quantizers.198 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.50e-02 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.down_proj_weight_quantizers.199 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.down_proj_weight_quantizers.200 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.54e-02 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.down_proj_weight_quantizers.201 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.35e-01 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.down_proj_weight_quantizers.202 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.down_proj_weight_quantizers.203 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.down_proj_weight_quantizers.204 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.54e-02 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.down_proj_weight_quantizers.205 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.down_proj_weight_quantizers.206 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.down_proj_weight_quantizers.207 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.down_proj_weight_quantizers.208 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.down_proj_weight_quantizers.209 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.down_proj_weight_quantizers.210 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.down_proj_weight_quantizers.211 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.down_proj_weight_quantizers.212 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.64e-01 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.down_proj_weight_quantizers.213 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.down_proj_weight_quantizers.214 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.down_proj_weight_quantizers.215 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.down_proj_weight_quantizers.216 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.down_proj_weight_quantizers.217 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.down_proj_weight_quantizers.218 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.down_proj_weight_quantizers.219 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.down_proj_weight_quantizers.220 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.down_proj_weight_quantizers.221 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.down_proj_weight_quantizers.222 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.down_proj_weight_quantizers.223 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.75e-01 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.down_proj_weight_quantizers.224 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.down_proj_weight_quantizers.225 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.50e-01 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.down_proj_weight_quantizers.226 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.down_proj_weight_quantizers.227 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.79e-01 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.down_proj_weight_quantizers.228 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.22e-01 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.down_proj_weight_quantizers.229 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.down_proj_weight_quantizers.230 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.down_proj_weight_quantizers.231 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.02e-01 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.down_proj_weight_quantizers.232 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.down_proj_weight_quantizers.233 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.54e-02 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.down_proj_weight_quantizers.234 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.down_proj_weight_quantizers.235 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.down_proj_weight_quantizers.236 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.down_proj_weight_quantizers.237 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.down_proj_weight_quantizers.238 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.down_proj_weight_quantizers.239 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.down_proj_weight_quantizers.240 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.down_proj_weight_quantizers.241 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.down_proj_weight_quantizers.242 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.down_proj_weight_quantizers.243 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.down_proj_weight_quantizers.244 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.down_proj_weight_quantizers.245 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.down_proj_weight_quantizers.246 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.down_proj_weight_quantizers.247 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.down_proj_weight_quantizers.248 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.33e-01 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.down_proj_weight_quantizers.249 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.down_proj_weight_quantizers.250 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.down_proj_weight_quantizers.251 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.down_proj_weight_quantizers.252 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.21e-01 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.down_proj_weight_quantizers.253 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.70e-01 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.down_proj_weight_quantizers.254 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.down_proj_weight_quantizers.255 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.59e-02 calibrator=MaxCalibrator quant) +model.layers.45.mlp.shared_experts.gate_proj.input_quantizer TensorQuantizer(disabled) +model.layers.45.mlp.shared_experts.gate_proj.output_quantizer TensorQuantizer(disabled) +model.layers.45.mlp.shared_experts.gate_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.45.mlp.shared_experts.up_proj.input_quantizer TensorQuantizer(disabled) +model.layers.45.mlp.shared_experts.up_proj.output_quantizer TensorQuantizer(disabled) +model.layers.45.mlp.shared_experts.up_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.45.mlp.shared_experts.down_proj.input_quantizer TensorQuantizer(disabled) +model.layers.45.mlp.shared_experts.down_proj.output_quantizer TensorQuantizer(disabled) +model.layers.45.mlp.shared_experts.down_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.46.self_attn.q_a_proj.input_quantizer TensorQuantizer(disabled) +model.layers.46.self_attn.q_a_proj.output_quantizer TensorQuantizer(disabled) +model.layers.46.self_attn.q_a_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.46.self_attn.q_b_proj.input_quantizer TensorQuantizer(disabled) +model.layers.46.self_attn.q_b_proj.output_quantizer TensorQuantizer(disabled) +model.layers.46.self_attn.q_b_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.46.self_attn.kv_a_proj_with_mqa.input_quantizer TensorQuantizer(disabled) +model.layers.46.self_attn.kv_a_proj_with_mqa.output_quantizer TensorQuantizer(disabled) +model.layers.46.self_attn.kv_a_proj_with_mqa.weight_quantizer TensorQuantizer(disabled) +model.layers.46.self_attn.kv_b_proj.input_quantizer TensorQuantizer(disabled) +model.layers.46.self_attn.kv_b_proj.output_quantizer TensorQuantizer(disabled) +model.layers.46.self_attn.kv_b_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.46.self_attn.o_proj.input_quantizer TensorQuantizer(disabled) +model.layers.46.self_attn.o_proj.output_quantizer TensorQuantizer(disabled) +model.layers.46.self_attn.o_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.46.self_attn.indexer.wq_b.input_quantizer TensorQuantizer(disabled) +model.layers.46.self_attn.indexer.wq_b.output_quantizer TensorQuantizer(disabled) +model.layers.46.self_attn.indexer.wq_b.weight_quantizer TensorQuantizer(disabled) +model.layers.46.self_attn.indexer.wk.input_quantizer TensorQuantizer(disabled) +model.layers.46.self_attn.indexer.wk.output_quantizer TensorQuantizer(disabled) +model.layers.46.self_attn.indexer.wk.weight_quantizer TensorQuantizer(disabled) +model.layers.46.self_attn.indexer.k_norm.input_quantizer TensorQuantizer(disabled) +model.layers.46.self_attn.indexer.k_norm.output_quantizer TensorQuantizer(disabled) +model.layers.46.self_attn.indexer.weights_proj.input_quantizer TensorQuantizer(disabled) +model.layers.46.self_attn.indexer.weights_proj.output_quantizer TensorQuantizer(disabled) +model.layers.46.self_attn.indexer.weights_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.46.self_attn.q_bmm_quantizer TensorQuantizer(disabled) +model.layers.46.self_attn.k_bmm_quantizer TensorQuantizer((4, 3) bit fake per-tensor amax=4.48e+02(const) calibrator=MaxCalibrator quant) +model.layers.46.self_attn.v_bmm_quantizer TensorQuantizer((4, 3) bit fake per-tensor amax=4.48e+02(const) calibrator=MaxCalibrator quant) +model.layers.46.self_attn.p_bmm_quantizer TensorQuantizer(disabled) +model.layers.46.mlp.experts.gate_up_proj_input_quantizer TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.30e+00 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.gate_up_proj_weight_quantizers.0 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.gate_up_proj_weight_quantizers.1 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.72e-01 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.gate_up_proj_weight_quantizers.2 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.gate_up_proj_weight_quantizers.3 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.27e-01 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.gate_up_proj_weight_quantizers.4 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.15e-01 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.gate_up_proj_weight_quantizers.5 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.gate_up_proj_weight_quantizers.6 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.21e-01 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.gate_up_proj_weight_quantizers.7 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.70e-01 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.gate_up_proj_weight_quantizers.8 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.01e-01 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.gate_up_proj_weight_quantizers.9 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.47e-01 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.gate_up_proj_weight_quantizers.10 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.13e-01 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.gate_up_proj_weight_quantizers.11 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.gate_up_proj_weight_quantizers.12 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.15e-01 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.gate_up_proj_weight_quantizers.13 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.gate_up_proj_weight_quantizers.14 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.gate_up_proj_weight_quantizers.15 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.64e-02 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.gate_up_proj_weight_quantizers.16 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.gate_up_proj_weight_quantizers.17 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.gate_up_proj_weight_quantizers.18 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.gate_up_proj_weight_quantizers.19 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.gate_up_proj_weight_quantizers.20 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.gate_up_proj_weight_quantizers.21 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.gate_up_proj_weight_quantizers.22 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.12e-01 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.gate_up_proj_weight_quantizers.23 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.30e-01 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.gate_up_proj_weight_quantizers.24 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.gate_up_proj_weight_quantizers.25 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.gate_up_proj_weight_quantizers.26 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.gate_up_proj_weight_quantizers.27 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.27e-01 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.gate_up_proj_weight_quantizers.28 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.27e-01 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.gate_up_proj_weight_quantizers.29 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.gate_up_proj_weight_quantizers.30 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.23e-01 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.gate_up_proj_weight_quantizers.31 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.gate_up_proj_weight_quantizers.32 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.20e-01 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.gate_up_proj_weight_quantizers.33 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.gate_up_proj_weight_quantizers.34 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.gate_up_proj_weight_quantizers.35 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.80e-01 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.gate_up_proj_weight_quantizers.36 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.66e-01 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.gate_up_proj_weight_quantizers.37 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.76e-01 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.gate_up_proj_weight_quantizers.38 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.14e-01 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.gate_up_proj_weight_quantizers.39 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.16e-01 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.gate_up_proj_weight_quantizers.40 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.52e-01 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.gate_up_proj_weight_quantizers.41 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.20e-01 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.gate_up_proj_weight_quantizers.42 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.gate_up_proj_weight_quantizers.43 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.81e-01 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.gate_up_proj_weight_quantizers.44 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.15e-01 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.gate_up_proj_weight_quantizers.45 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.gate_up_proj_weight_quantizers.46 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.15e-01 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.gate_up_proj_weight_quantizers.47 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.28e-01 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.gate_up_proj_weight_quantizers.48 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.23e-01 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.gate_up_proj_weight_quantizers.49 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.gate_up_proj_weight_quantizers.50 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.gate_up_proj_weight_quantizers.51 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.20e-01 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.gate_up_proj_weight_quantizers.52 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.gate_up_proj_weight_quantizers.53 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.16e-01 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.gate_up_proj_weight_quantizers.54 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.40e-01 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.gate_up_proj_weight_quantizers.55 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.25e-01 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.gate_up_proj_weight_quantizers.56 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.30e-01 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.gate_up_proj_weight_quantizers.57 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.gate_up_proj_weight_quantizers.58 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.gate_up_proj_weight_quantizers.59 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.56e-01 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.gate_up_proj_weight_quantizers.60 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.46e-01 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.gate_up_proj_weight_quantizers.61 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.gate_up_proj_weight_quantizers.62 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.18e-01 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.gate_up_proj_weight_quantizers.63 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.31e-01 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.gate_up_proj_weight_quantizers.64 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.30e-01 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.gate_up_proj_weight_quantizers.65 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.16e-01 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.gate_up_proj_weight_quantizers.66 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.58e-01 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.gate_up_proj_weight_quantizers.67 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.gate_up_proj_weight_quantizers.68 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.gate_up_proj_weight_quantizers.69 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.gate_up_proj_weight_quantizers.70 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.gate_up_proj_weight_quantizers.71 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.gate_up_proj_weight_quantizers.72 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.gate_up_proj_weight_quantizers.73 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.09e-01 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.gate_up_proj_weight_quantizers.74 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.gate_up_proj_weight_quantizers.75 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.36e-01 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.gate_up_proj_weight_quantizers.76 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.gate_up_proj_weight_quantizers.77 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.16e-01 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.gate_up_proj_weight_quantizers.78 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.14e-01 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.gate_up_proj_weight_quantizers.79 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.22e-01 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.gate_up_proj_weight_quantizers.80 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.gate_up_proj_weight_quantizers.81 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.17e-01 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.gate_up_proj_weight_quantizers.82 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.24e-01 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.gate_up_proj_weight_quantizers.83 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.19e-01 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.gate_up_proj_weight_quantizers.84 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.gate_up_proj_weight_quantizers.85 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.26e-01 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.gate_up_proj_weight_quantizers.86 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.43e-01 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.gate_up_proj_weight_quantizers.87 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.29e-01 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.gate_up_proj_weight_quantizers.88 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.gate_up_proj_weight_quantizers.89 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.gate_up_proj_weight_quantizers.90 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.31e-01 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.gate_up_proj_weight_quantizers.91 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.gate_up_proj_weight_quantizers.92 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.73e-01 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.gate_up_proj_weight_quantizers.93 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.gate_up_proj_weight_quantizers.94 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.gate_up_proj_weight_quantizers.95 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.gate_up_proj_weight_quantizers.96 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.28e-01 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.gate_up_proj_weight_quantizers.97 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.gate_up_proj_weight_quantizers.98 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.59e-01 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.gate_up_proj_weight_quantizers.99 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.gate_up_proj_weight_quantizers.100 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.gate_up_proj_weight_quantizers.101 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.32e-01 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.gate_up_proj_weight_quantizers.102 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.gate_up_proj_weight_quantizers.103 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.gate_up_proj_weight_quantizers.104 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.gate_up_proj_weight_quantizers.105 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.09e-01 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.gate_up_proj_weight_quantizers.106 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.44e-01 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.gate_up_proj_weight_quantizers.107 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.gate_up_proj_weight_quantizers.108 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.gate_up_proj_weight_quantizers.109 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.gate_up_proj_weight_quantizers.110 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.gate_up_proj_weight_quantizers.111 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.74e-01 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.gate_up_proj_weight_quantizers.112 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.32e-01 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.gate_up_proj_weight_quantizers.113 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.gate_up_proj_weight_quantizers.114 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.21e-01 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.gate_up_proj_weight_quantizers.115 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.09e-01 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.gate_up_proj_weight_quantizers.116 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.70e-01 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.gate_up_proj_weight_quantizers.117 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.gate_up_proj_weight_quantizers.118 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.gate_up_proj_weight_quantizers.119 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.14e-01 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.gate_up_proj_weight_quantizers.120 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.gate_up_proj_weight_quantizers.121 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.29e-01 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.gate_up_proj_weight_quantizers.122 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.35e-01 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.gate_up_proj_weight_quantizers.123 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.gate_up_proj_weight_quantizers.124 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.80e-01 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.gate_up_proj_weight_quantizers.125 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.gate_up_proj_weight_quantizers.126 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.20e-01 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.gate_up_proj_weight_quantizers.127 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.gate_up_proj_weight_quantizers.128 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.gate_up_proj_weight_quantizers.129 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.gate_up_proj_weight_quantizers.130 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.gate_up_proj_weight_quantizers.131 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.28e-01 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.gate_up_proj_weight_quantizers.132 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.32e-01 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.gate_up_proj_weight_quantizers.133 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.gate_up_proj_weight_quantizers.134 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.47e-01 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.gate_up_proj_weight_quantizers.135 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.gate_up_proj_weight_quantizers.136 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.66e-01 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.gate_up_proj_weight_quantizers.137 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.gate_up_proj_weight_quantizers.138 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.18e-01 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.gate_up_proj_weight_quantizers.139 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.19e-01 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.gate_up_proj_weight_quantizers.140 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.gate_up_proj_weight_quantizers.141 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.45e-01 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.gate_up_proj_weight_quantizers.142 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.40e-01 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.gate_up_proj_weight_quantizers.143 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.gate_up_proj_weight_quantizers.144 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.27e-01 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.gate_up_proj_weight_quantizers.145 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.gate_up_proj_weight_quantizers.146 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.gate_up_proj_weight_quantizers.147 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.gate_up_proj_weight_quantizers.148 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.23e-01 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.gate_up_proj_weight_quantizers.149 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.gate_up_proj_weight_quantizers.150 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.gate_up_proj_weight_quantizers.151 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.gate_up_proj_weight_quantizers.152 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.28e-01 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.gate_up_proj_weight_quantizers.153 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.gate_up_proj_weight_quantizers.154 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.98e-01 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.gate_up_proj_weight_quantizers.155 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.31e-01 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.gate_up_proj_weight_quantizers.156 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.gate_up_proj_weight_quantizers.157 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.37e-01 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.gate_up_proj_weight_quantizers.158 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.gate_up_proj_weight_quantizers.159 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.gate_up_proj_weight_quantizers.160 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.gate_up_proj_weight_quantizers.161 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.gate_up_proj_weight_quantizers.162 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.36e-01 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.gate_up_proj_weight_quantizers.163 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.88e-01 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.gate_up_proj_weight_quantizers.164 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.14e-01 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.gate_up_proj_weight_quantizers.165 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.18e-01 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.gate_up_proj_weight_quantizers.166 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.gate_up_proj_weight_quantizers.167 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.73e-01 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.gate_up_proj_weight_quantizers.168 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.gate_up_proj_weight_quantizers.169 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.gate_up_proj_weight_quantizers.170 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.42e-01 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.gate_up_proj_weight_quantizers.171 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.30e-01 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.gate_up_proj_weight_quantizers.172 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.23e-01 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.gate_up_proj_weight_quantizers.173 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.52e-01 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.gate_up_proj_weight_quantizers.174 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.27e-01 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.gate_up_proj_weight_quantizers.175 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.52e-01 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.gate_up_proj_weight_quantizers.176 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.gate_up_proj_weight_quantizers.177 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.gate_up_proj_weight_quantizers.178 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.25e-01 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.gate_up_proj_weight_quantizers.179 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.13e-01 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.gate_up_proj_weight_quantizers.180 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.20e-01 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.gate_up_proj_weight_quantizers.181 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.14e-01 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.gate_up_proj_weight_quantizers.182 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.gate_up_proj_weight_quantizers.183 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.gate_up_proj_weight_quantizers.184 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.60e-01 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.gate_up_proj_weight_quantizers.185 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.12e-01 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.gate_up_proj_weight_quantizers.186 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.21e-01 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.gate_up_proj_weight_quantizers.187 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.gate_up_proj_weight_quantizers.188 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.13e-01 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.gate_up_proj_weight_quantizers.189 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.24e-01 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.gate_up_proj_weight_quantizers.190 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.gate_up_proj_weight_quantizers.191 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.gate_up_proj_weight_quantizers.192 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.gate_up_proj_weight_quantizers.193 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.gate_up_proj_weight_quantizers.194 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.gate_up_proj_weight_quantizers.195 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.15e-01 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.gate_up_proj_weight_quantizers.196 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.gate_up_proj_weight_quantizers.197 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.95e-01 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.gate_up_proj_weight_quantizers.198 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.38e-01 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.gate_up_proj_weight_quantizers.199 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.52e-01 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.gate_up_proj_weight_quantizers.200 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.gate_up_proj_weight_quantizers.201 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.gate_up_proj_weight_quantizers.202 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.gate_up_proj_weight_quantizers.203 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.gate_up_proj_weight_quantizers.204 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.gate_up_proj_weight_quantizers.205 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.gate_up_proj_weight_quantizers.206 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.47e-01 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.gate_up_proj_weight_quantizers.207 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.gate_up_proj_weight_quantizers.208 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.gate_up_proj_weight_quantizers.209 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.gate_up_proj_weight_quantizers.210 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.gate_up_proj_weight_quantizers.211 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.gate_up_proj_weight_quantizers.212 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.61e-01 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.gate_up_proj_weight_quantizers.213 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.gate_up_proj_weight_quantizers.214 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.gate_up_proj_weight_quantizers.215 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.gate_up_proj_weight_quantizers.216 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.gate_up_proj_weight_quantizers.217 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.gate_up_proj_weight_quantizers.218 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.gate_up_proj_weight_quantizers.219 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.gate_up_proj_weight_quantizers.220 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.43e-01 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.gate_up_proj_weight_quantizers.221 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.gate_up_proj_weight_quantizers.222 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.gate_up_proj_weight_quantizers.223 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.37e-01 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.gate_up_proj_weight_quantizers.224 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.gate_up_proj_weight_quantizers.225 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.29e-01 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.gate_up_proj_weight_quantizers.226 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.gate_up_proj_weight_quantizers.227 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.16e-01 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.gate_up_proj_weight_quantizers.228 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.30e-01 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.gate_up_proj_weight_quantizers.229 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.64e-01 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.gate_up_proj_weight_quantizers.230 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.gate_up_proj_weight_quantizers.231 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.44e-01 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.gate_up_proj_weight_quantizers.232 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.gate_up_proj_weight_quantizers.233 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.23e-01 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.gate_up_proj_weight_quantizers.234 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.gate_up_proj_weight_quantizers.235 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.gate_up_proj_weight_quantizers.236 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.gate_up_proj_weight_quantizers.237 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.gate_up_proj_weight_quantizers.238 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.gate_up_proj_weight_quantizers.239 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.gate_up_proj_weight_quantizers.240 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.47e-01 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.gate_up_proj_weight_quantizers.241 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.gate_up_proj_weight_quantizers.242 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.gate_up_proj_weight_quantizers.243 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.gate_up_proj_weight_quantizers.244 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.gate_up_proj_weight_quantizers.245 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.gate_up_proj_weight_quantizers.246 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.gate_up_proj_weight_quantizers.247 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.25e-01 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.gate_up_proj_weight_quantizers.248 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.gate_up_proj_weight_quantizers.249 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.gate_up_proj_weight_quantizers.250 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.23e-01 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.gate_up_proj_weight_quantizers.251 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.gate_up_proj_weight_quantizers.252 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.gate_up_proj_weight_quantizers.253 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.07e-01 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.gate_up_proj_weight_quantizers.254 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.gate_up_proj_weight_quantizers.255 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.down_proj_input_quantizer TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.98e+01 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.down_proj_weight_quantizers.0 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.down_proj_weight_quantizers.1 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.down_proj_weight_quantizers.2 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.down_proj_weight_quantizers.3 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.48e-01 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.down_proj_weight_quantizers.4 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.down_proj_weight_quantizers.5 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.down_proj_weight_quantizers.6 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.down_proj_weight_quantizers.7 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.down_proj_weight_quantizers.8 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=4.20e-01 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.down_proj_weight_quantizers.9 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.down_proj_weight_quantizers.10 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.17e-01 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.down_proj_weight_quantizers.11 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.down_proj_weight_quantizers.12 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.down_proj_weight_quantizers.13 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.44e-01 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.down_proj_weight_quantizers.14 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.down_proj_weight_quantizers.15 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.down_proj_weight_quantizers.16 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.down_proj_weight_quantizers.17 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.down_proj_weight_quantizers.18 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.down_proj_weight_quantizers.19 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.53e-01 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.down_proj_weight_quantizers.20 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.down_proj_weight_quantizers.21 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.down_proj_weight_quantizers.22 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.down_proj_weight_quantizers.23 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.down_proj_weight_quantizers.24 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.59e-02 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.down_proj_weight_quantizers.25 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.down_proj_weight_quantizers.26 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.down_proj_weight_quantizers.27 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.50e-01 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.down_proj_weight_quantizers.28 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.down_proj_weight_quantizers.29 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.down_proj_weight_quantizers.30 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.23e-01 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.down_proj_weight_quantizers.31 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.down_proj_weight_quantizers.32 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.down_proj_weight_quantizers.33 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.59e-02 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.down_proj_weight_quantizers.34 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.down_proj_weight_quantizers.35 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.44e-01 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.down_proj_weight_quantizers.36 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.down_proj_weight_quantizers.37 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.50e-01 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.down_proj_weight_quantizers.38 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.down_proj_weight_quantizers.39 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.59e-02 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.down_proj_weight_quantizers.40 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.down_proj_weight_quantizers.41 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.down_proj_weight_quantizers.42 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.down_proj_weight_quantizers.43 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.64e-01 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.down_proj_weight_quantizers.44 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.down_proj_weight_quantizers.45 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.down_proj_weight_quantizers.46 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.94e-01 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.down_proj_weight_quantizers.47 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.down_proj_weight_quantizers.48 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.45e-01 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.down_proj_weight_quantizers.49 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.54e-02 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.down_proj_weight_quantizers.50 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.40e-02 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.down_proj_weight_quantizers.51 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.down_proj_weight_quantizers.52 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.down_proj_weight_quantizers.53 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.down_proj_weight_quantizers.54 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.45e-02 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.down_proj_weight_quantizers.55 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.59e-02 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.down_proj_weight_quantizers.56 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.down_proj_weight_quantizers.57 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.down_proj_weight_quantizers.58 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.down_proj_weight_quantizers.59 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.down_proj_weight_quantizers.60 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.down_proj_weight_quantizers.61 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.50e-02 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.down_proj_weight_quantizers.62 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.down_proj_weight_quantizers.63 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.down_proj_weight_quantizers.64 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.54e-01 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.down_proj_weight_quantizers.65 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.down_proj_weight_quantizers.66 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.77e-01 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.down_proj_weight_quantizers.67 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.down_proj_weight_quantizers.68 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.down_proj_weight_quantizers.69 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.down_proj_weight_quantizers.70 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.down_proj_weight_quantizers.71 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.down_proj_weight_quantizers.72 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.down_proj_weight_quantizers.73 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.19e-01 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.down_proj_weight_quantizers.74 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.down_proj_weight_quantizers.75 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.43e-01 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.down_proj_weight_quantizers.76 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.down_proj_weight_quantizers.77 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.49e-01 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.down_proj_weight_quantizers.78 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.75e-01 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.down_proj_weight_quantizers.79 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.down_proj_weight_quantizers.80 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.down_proj_weight_quantizers.81 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.down_proj_weight_quantizers.82 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.20e-01 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.down_proj_weight_quantizers.83 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.64e-02 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.down_proj_weight_quantizers.84 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.down_proj_weight_quantizers.85 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.down_proj_weight_quantizers.86 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.20e-01 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.down_proj_weight_quantizers.87 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.down_proj_weight_quantizers.88 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.down_proj_weight_quantizers.89 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.down_proj_weight_quantizers.90 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.down_proj_weight_quantizers.91 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.down_proj_weight_quantizers.92 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.44e-01 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.down_proj_weight_quantizers.93 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.54e-02 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.down_proj_weight_quantizers.94 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.83e-01 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.down_proj_weight_quantizers.95 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.down_proj_weight_quantizers.96 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.31e-01 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.down_proj_weight_quantizers.97 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.down_proj_weight_quantizers.98 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.66e-01 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.down_proj_weight_quantizers.99 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.40e-01 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.down_proj_weight_quantizers.100 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.down_proj_weight_quantizers.101 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.down_proj_weight_quantizers.102 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.15e-02 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.down_proj_weight_quantizers.103 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.down_proj_weight_quantizers.104 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.50e-02 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.down_proj_weight_quantizers.105 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.35e-01 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.down_proj_weight_quantizers.106 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.down_proj_weight_quantizers.107 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.down_proj_weight_quantizers.108 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.64e-02 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.down_proj_weight_quantizers.109 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.down_proj_weight_quantizers.110 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.59e-02 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.down_proj_weight_quantizers.111 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.67e-01 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.down_proj_weight_quantizers.112 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.17e-01 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.down_proj_weight_quantizers.113 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.down_proj_weight_quantizers.114 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.down_proj_weight_quantizers.115 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.21e-01 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.down_proj_weight_quantizers.116 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.14e-01 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.down_proj_weight_quantizers.117 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.down_proj_weight_quantizers.118 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.45e-02 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.down_proj_weight_quantizers.119 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.down_proj_weight_quantizers.120 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.21e-01 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.down_proj_weight_quantizers.121 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.down_proj_weight_quantizers.122 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.down_proj_weight_quantizers.123 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.59e-02 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.down_proj_weight_quantizers.124 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.down_proj_weight_quantizers.125 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.down_proj_weight_quantizers.126 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.down_proj_weight_quantizers.127 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.25e-01 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.down_proj_weight_quantizers.128 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.54e-01 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.down_proj_weight_quantizers.129 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.down_proj_weight_quantizers.130 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.down_proj_weight_quantizers.131 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.43e-01 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.down_proj_weight_quantizers.132 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.down_proj_weight_quantizers.133 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.down_proj_weight_quantizers.134 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.50e-01 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.down_proj_weight_quantizers.135 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.down_proj_weight_quantizers.136 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.37e-01 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.down_proj_weight_quantizers.137 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.down_proj_weight_quantizers.138 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.down_proj_weight_quantizers.139 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.down_proj_weight_quantizers.140 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.33e-01 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.down_proj_weight_quantizers.141 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.54e-02 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.down_proj_weight_quantizers.142 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.down_proj_weight_quantizers.143 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.59e-02 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.down_proj_weight_quantizers.144 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.down_proj_weight_quantizers.145 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.down_proj_weight_quantizers.146 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.down_proj_weight_quantizers.147 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.down_proj_weight_quantizers.148 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.35e-01 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.down_proj_weight_quantizers.149 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.down_proj_weight_quantizers.150 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.down_proj_weight_quantizers.151 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.down_proj_weight_quantizers.152 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.13e-01 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.down_proj_weight_quantizers.153 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.down_proj_weight_quantizers.154 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.56e-01 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.down_proj_weight_quantizers.155 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.54e-02 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.down_proj_weight_quantizers.156 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.down_proj_weight_quantizers.157 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.20e-01 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.down_proj_weight_quantizers.158 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.94e-01 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.down_proj_weight_quantizers.159 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.down_proj_weight_quantizers.160 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.59e-02 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.down_proj_weight_quantizers.161 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.down_proj_weight_quantizers.162 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.down_proj_weight_quantizers.163 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.49e-01 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.down_proj_weight_quantizers.164 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.down_proj_weight_quantizers.165 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.down_proj_weight_quantizers.166 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.45e-02 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.down_proj_weight_quantizers.167 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.down_proj_weight_quantizers.168 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.down_proj_weight_quantizers.169 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.down_proj_weight_quantizers.170 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.down_proj_weight_quantizers.171 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.down_proj_weight_quantizers.172 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.12e-01 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.down_proj_weight_quantizers.173 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.down_proj_weight_quantizers.174 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.down_proj_weight_quantizers.175 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.62e-01 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.down_proj_weight_quantizers.176 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.70e-01 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.down_proj_weight_quantizers.177 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.down_proj_weight_quantizers.178 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.29e-01 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.down_proj_weight_quantizers.179 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.45e-02 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.down_proj_weight_quantizers.180 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.29e-01 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.down_proj_weight_quantizers.181 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.down_proj_weight_quantizers.182 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.down_proj_weight_quantizers.183 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.down_proj_weight_quantizers.184 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.down_proj_weight_quantizers.185 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.down_proj_weight_quantizers.186 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.90e-01 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.down_proj_weight_quantizers.187 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.down_proj_weight_quantizers.188 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.down_proj_weight_quantizers.189 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.down_proj_weight_quantizers.190 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.59e-02 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.down_proj_weight_quantizers.191 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.45e-02 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.down_proj_weight_quantizers.192 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.down_proj_weight_quantizers.193 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.down_proj_weight_quantizers.194 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.down_proj_weight_quantizers.195 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.down_proj_weight_quantizers.196 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.down_proj_weight_quantizers.197 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.down_proj_weight_quantizers.198 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.down_proj_weight_quantizers.199 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.64e-02 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.down_proj_weight_quantizers.200 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.down_proj_weight_quantizers.201 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.down_proj_weight_quantizers.202 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.down_proj_weight_quantizers.203 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.down_proj_weight_quantizers.204 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.down_proj_weight_quantizers.205 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.down_proj_weight_quantizers.206 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.down_proj_weight_quantizers.207 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.down_proj_weight_quantizers.208 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.down_proj_weight_quantizers.209 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.down_proj_weight_quantizers.210 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.down_proj_weight_quantizers.211 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.down_proj_weight_quantizers.212 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.12e-01 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.down_proj_weight_quantizers.213 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.down_proj_weight_quantizers.214 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.down_proj_weight_quantizers.215 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.down_proj_weight_quantizers.216 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.down_proj_weight_quantizers.217 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.down_proj_weight_quantizers.218 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.down_proj_weight_quantizers.219 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.down_proj_weight_quantizers.220 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.down_proj_weight_quantizers.221 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.54e-01 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.down_proj_weight_quantizers.222 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.down_proj_weight_quantizers.223 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.23e-01 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.down_proj_weight_quantizers.224 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.down_proj_weight_quantizers.225 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.down_proj_weight_quantizers.226 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.down_proj_weight_quantizers.227 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.76e-01 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.down_proj_weight_quantizers.228 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.down_proj_weight_quantizers.229 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.29e-01 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.down_proj_weight_quantizers.230 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.down_proj_weight_quantizers.231 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.65e-01 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.down_proj_weight_quantizers.232 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.down_proj_weight_quantizers.233 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.down_proj_weight_quantizers.234 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.30e-02 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.down_proj_weight_quantizers.235 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.down_proj_weight_quantizers.236 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.down_proj_weight_quantizers.237 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.55e-01 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.down_proj_weight_quantizers.238 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.45e-02 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.down_proj_weight_quantizers.239 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.down_proj_weight_quantizers.240 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.down_proj_weight_quantizers.241 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.down_proj_weight_quantizers.242 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.down_proj_weight_quantizers.243 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.down_proj_weight_quantizers.244 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.down_proj_weight_quantizers.245 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.down_proj_weight_quantizers.246 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.down_proj_weight_quantizers.247 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.down_proj_weight_quantizers.248 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.35e-02 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.down_proj_weight_quantizers.249 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.down_proj_weight_quantizers.250 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.down_proj_weight_quantizers.251 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.down_proj_weight_quantizers.252 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.down_proj_weight_quantizers.253 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.46e-01 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.down_proj_weight_quantizers.254 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.down_proj_weight_quantizers.255 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.46.mlp.shared_experts.gate_proj.input_quantizer TensorQuantizer(disabled) +model.layers.46.mlp.shared_experts.gate_proj.output_quantizer TensorQuantizer(disabled) +model.layers.46.mlp.shared_experts.gate_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.46.mlp.shared_experts.up_proj.input_quantizer TensorQuantizer(disabled) +model.layers.46.mlp.shared_experts.up_proj.output_quantizer TensorQuantizer(disabled) +model.layers.46.mlp.shared_experts.up_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.46.mlp.shared_experts.down_proj.input_quantizer TensorQuantizer(disabled) +model.layers.46.mlp.shared_experts.down_proj.output_quantizer TensorQuantizer(disabled) +model.layers.46.mlp.shared_experts.down_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.47.self_attn.q_a_proj.input_quantizer TensorQuantizer(disabled) +model.layers.47.self_attn.q_a_proj.output_quantizer TensorQuantizer(disabled) +model.layers.47.self_attn.q_a_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.47.self_attn.q_b_proj.input_quantizer TensorQuantizer(disabled) +model.layers.47.self_attn.q_b_proj.output_quantizer TensorQuantizer(disabled) +model.layers.47.self_attn.q_b_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.47.self_attn.kv_a_proj_with_mqa.input_quantizer TensorQuantizer(disabled) +model.layers.47.self_attn.kv_a_proj_with_mqa.output_quantizer TensorQuantizer(disabled) +model.layers.47.self_attn.kv_a_proj_with_mqa.weight_quantizer TensorQuantizer(disabled) +model.layers.47.self_attn.kv_b_proj.input_quantizer TensorQuantizer(disabled) +model.layers.47.self_attn.kv_b_proj.output_quantizer TensorQuantizer(disabled) +model.layers.47.self_attn.kv_b_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.47.self_attn.o_proj.input_quantizer TensorQuantizer(disabled) +model.layers.47.self_attn.o_proj.output_quantizer TensorQuantizer(disabled) +model.layers.47.self_attn.o_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.47.self_attn.q_bmm_quantizer TensorQuantizer(disabled) +model.layers.47.self_attn.k_bmm_quantizer TensorQuantizer((4, 3) bit fake per-tensor amax=4.48e+02(const) calibrator=MaxCalibrator quant) +model.layers.47.self_attn.v_bmm_quantizer TensorQuantizer((4, 3) bit fake per-tensor amax=4.48e+02(const) calibrator=MaxCalibrator quant) +model.layers.47.self_attn.p_bmm_quantizer TensorQuantizer(disabled) +model.layers.47.mlp.experts.gate_up_proj_input_quantizer TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.41e+00 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.gate_up_proj_weight_quantizers.0 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.gate_up_proj_weight_quantizers.1 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.gate_up_proj_weight_quantizers.2 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.16e-01 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.gate_up_proj_weight_quantizers.3 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.gate_up_proj_weight_quantizers.4 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.29e-01 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.gate_up_proj_weight_quantizers.5 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.gate_up_proj_weight_quantizers.6 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.gate_up_proj_weight_quantizers.7 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.gate_up_proj_weight_quantizers.8 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.gate_up_proj_weight_quantizers.9 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.gate_up_proj_weight_quantizers.10 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.gate_up_proj_weight_quantizers.11 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.gate_up_proj_weight_quantizers.12 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.gate_up_proj_weight_quantizers.13 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.15e-01 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.gate_up_proj_weight_quantizers.14 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.gate_up_proj_weight_quantizers.15 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.39e-01 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.gate_up_proj_weight_quantizers.16 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.16e-01 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.gate_up_proj_weight_quantizers.17 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.17e-01 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.gate_up_proj_weight_quantizers.18 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.51e-01 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.gate_up_proj_weight_quantizers.19 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.gate_up_proj_weight_quantizers.20 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.28e-01 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.gate_up_proj_weight_quantizers.21 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.gate_up_proj_weight_quantizers.22 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.gate_up_proj_weight_quantizers.23 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.gate_up_proj_weight_quantizers.24 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.24e-01 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.gate_up_proj_weight_quantizers.25 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.gate_up_proj_weight_quantizers.26 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.36e-01 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.gate_up_proj_weight_quantizers.27 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.24e-01 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.gate_up_proj_weight_quantizers.28 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.gate_up_proj_weight_quantizers.29 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.20e-01 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.gate_up_proj_weight_quantizers.30 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.14e-01 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.gate_up_proj_weight_quantizers.31 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.47e-01 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.gate_up_proj_weight_quantizers.32 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.gate_up_proj_weight_quantizers.33 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.gate_up_proj_weight_quantizers.34 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.gate_up_proj_weight_quantizers.35 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.gate_up_proj_weight_quantizers.36 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.29e-01 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.gate_up_proj_weight_quantizers.37 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.16e-01 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.gate_up_proj_weight_quantizers.38 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.21e-01 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.gate_up_proj_weight_quantizers.39 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.25e-01 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.gate_up_proj_weight_quantizers.40 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.17e-01 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.gate_up_proj_weight_quantizers.41 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.13e-01 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.gate_up_proj_weight_quantizers.42 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.93e-01 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.gate_up_proj_weight_quantizers.43 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.gate_up_proj_weight_quantizers.44 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.gate_up_proj_weight_quantizers.45 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.gate_up_proj_weight_quantizers.46 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.gate_up_proj_weight_quantizers.47 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.gate_up_proj_weight_quantizers.48 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.gate_up_proj_weight_quantizers.49 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.gate_up_proj_weight_quantizers.50 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.gate_up_proj_weight_quantizers.51 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.18e-01 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.gate_up_proj_weight_quantizers.52 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.gate_up_proj_weight_quantizers.53 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.20e-01 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.gate_up_proj_weight_quantizers.54 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.gate_up_proj_weight_quantizers.55 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.gate_up_proj_weight_quantizers.56 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.gate_up_proj_weight_quantizers.57 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.22e-01 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.gate_up_proj_weight_quantizers.58 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.28e-01 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.gate_up_proj_weight_quantizers.59 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.gate_up_proj_weight_quantizers.60 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.gate_up_proj_weight_quantizers.61 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.22e-01 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.gate_up_proj_weight_quantizers.62 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.gate_up_proj_weight_quantizers.63 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.gate_up_proj_weight_quantizers.64 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.gate_up_proj_weight_quantizers.65 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.gate_up_proj_weight_quantizers.66 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.gate_up_proj_weight_quantizers.67 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.43e-01 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.gate_up_proj_weight_quantizers.68 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.gate_up_proj_weight_quantizers.69 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.21e-01 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.gate_up_proj_weight_quantizers.70 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.gate_up_proj_weight_quantizers.71 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.24e-01 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.gate_up_proj_weight_quantizers.72 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.gate_up_proj_weight_quantizers.73 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.gate_up_proj_weight_quantizers.74 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.52e-01 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.gate_up_proj_weight_quantizers.75 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.75e-01 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.gate_up_proj_weight_quantizers.76 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.gate_up_proj_weight_quantizers.77 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.gate_up_proj_weight_quantizers.78 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.gate_up_proj_weight_quantizers.79 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.gate_up_proj_weight_quantizers.80 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.46e-01 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.gate_up_proj_weight_quantizers.81 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.14e-01 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.gate_up_proj_weight_quantizers.82 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.20e-01 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.gate_up_proj_weight_quantizers.83 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.gate_up_proj_weight_quantizers.84 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.16e-01 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.gate_up_proj_weight_quantizers.85 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.14e-01 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.gate_up_proj_weight_quantizers.86 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.34e-01 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.gate_up_proj_weight_quantizers.87 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.gate_up_proj_weight_quantizers.88 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.24e-01 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.gate_up_proj_weight_quantizers.89 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.gate_up_proj_weight_quantizers.90 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.gate_up_proj_weight_quantizers.91 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.gate_up_proj_weight_quantizers.92 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.39e-01 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.gate_up_proj_weight_quantizers.93 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.gate_up_proj_weight_quantizers.94 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.gate_up_proj_weight_quantizers.95 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.gate_up_proj_weight_quantizers.96 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.20e-01 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.gate_up_proj_weight_quantizers.97 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.gate_up_proj_weight_quantizers.98 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.gate_up_proj_weight_quantizers.99 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.gate_up_proj_weight_quantizers.100 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.05e-01 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.gate_up_proj_weight_quantizers.101 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.93e-01 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.gate_up_proj_weight_quantizers.102 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.gate_up_proj_weight_quantizers.103 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.gate_up_proj_weight_quantizers.104 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.26e-01 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.gate_up_proj_weight_quantizers.105 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.gate_up_proj_weight_quantizers.106 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.gate_up_proj_weight_quantizers.107 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.52e-01 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.gate_up_proj_weight_quantizers.108 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.19e-01 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.gate_up_proj_weight_quantizers.109 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.71e-01 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.gate_up_proj_weight_quantizers.110 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.44e-01 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.gate_up_proj_weight_quantizers.111 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.gate_up_proj_weight_quantizers.112 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.27e-01 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.gate_up_proj_weight_quantizers.113 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.gate_up_proj_weight_quantizers.114 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.gate_up_proj_weight_quantizers.115 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.79e-01 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.gate_up_proj_weight_quantizers.116 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.20e-01 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.gate_up_proj_weight_quantizers.117 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.gate_up_proj_weight_quantizers.118 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.22e-01 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.gate_up_proj_weight_quantizers.119 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.gate_up_proj_weight_quantizers.120 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.gate_up_proj_weight_quantizers.121 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.gate_up_proj_weight_quantizers.122 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.gate_up_proj_weight_quantizers.123 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.gate_up_proj_weight_quantizers.124 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.gate_up_proj_weight_quantizers.125 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.gate_up_proj_weight_quantizers.126 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.75e-01 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.gate_up_proj_weight_quantizers.127 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.gate_up_proj_weight_quantizers.128 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.14e-01 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.gate_up_proj_weight_quantizers.129 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.gate_up_proj_weight_quantizers.130 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.gate_up_proj_weight_quantizers.131 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.gate_up_proj_weight_quantizers.132 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.40e-01 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.gate_up_proj_weight_quantizers.133 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.gate_up_proj_weight_quantizers.134 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.gate_up_proj_weight_quantizers.135 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.33e-01 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.gate_up_proj_weight_quantizers.136 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.43e-01 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.gate_up_proj_weight_quantizers.137 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.01e-01 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.gate_up_proj_weight_quantizers.138 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.21e-01 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.gate_up_proj_weight_quantizers.139 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.15e-01 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.gate_up_proj_weight_quantizers.140 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.gate_up_proj_weight_quantizers.141 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.gate_up_proj_weight_quantizers.142 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.gate_up_proj_weight_quantizers.143 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.24e-01 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.gate_up_proj_weight_quantizers.144 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.18e-01 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.gate_up_proj_weight_quantizers.145 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.62e-01 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.gate_up_proj_weight_quantizers.146 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.55e-01 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.gate_up_proj_weight_quantizers.147 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.gate_up_proj_weight_quantizers.148 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.gate_up_proj_weight_quantizers.149 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.54e-01 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.gate_up_proj_weight_quantizers.150 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.16e-01 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.gate_up_proj_weight_quantizers.151 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.63e-01 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.gate_up_proj_weight_quantizers.152 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.gate_up_proj_weight_quantizers.153 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.32e-01 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.gate_up_proj_weight_quantizers.154 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.gate_up_proj_weight_quantizers.155 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.14e-01 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.gate_up_proj_weight_quantizers.156 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.73e-01 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.gate_up_proj_weight_quantizers.157 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.gate_up_proj_weight_quantizers.158 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.42e-01 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.gate_up_proj_weight_quantizers.159 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.96e-01 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.gate_up_proj_weight_quantizers.160 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.gate_up_proj_weight_quantizers.161 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.gate_up_proj_weight_quantizers.162 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.76e-01 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.gate_up_proj_weight_quantizers.163 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.gate_up_proj_weight_quantizers.164 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.49e-01 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.gate_up_proj_weight_quantizers.165 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.15e-01 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.gate_up_proj_weight_quantizers.166 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.gate_up_proj_weight_quantizers.167 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.gate_up_proj_weight_quantizers.168 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.18e-01 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.gate_up_proj_weight_quantizers.169 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.gate_up_proj_weight_quantizers.170 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.gate_up_proj_weight_quantizers.171 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.gate_up_proj_weight_quantizers.172 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.gate_up_proj_weight_quantizers.173 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.gate_up_proj_weight_quantizers.174 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.31e-01 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.gate_up_proj_weight_quantizers.175 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.gate_up_proj_weight_quantizers.176 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.gate_up_proj_weight_quantizers.177 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.gate_up_proj_weight_quantizers.178 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.30e-01 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.gate_up_proj_weight_quantizers.179 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.gate_up_proj_weight_quantizers.180 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.16e-01 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.gate_up_proj_weight_quantizers.181 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.gate_up_proj_weight_quantizers.182 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.24e-01 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.gate_up_proj_weight_quantizers.183 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.gate_up_proj_weight_quantizers.184 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.19e-01 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.gate_up_proj_weight_quantizers.185 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.gate_up_proj_weight_quantizers.186 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.21e-01 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.gate_up_proj_weight_quantizers.187 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.gate_up_proj_weight_quantizers.188 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.27e-01 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.gate_up_proj_weight_quantizers.189 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.16e-01 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.gate_up_proj_weight_quantizers.190 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.13e-01 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.gate_up_proj_weight_quantizers.191 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.94e-01 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.gate_up_proj_weight_quantizers.192 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.gate_up_proj_weight_quantizers.193 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.89e-01 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.gate_up_proj_weight_quantizers.194 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.50e-01 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.gate_up_proj_weight_quantizers.195 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.gate_up_proj_weight_quantizers.196 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.gate_up_proj_weight_quantizers.197 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.gate_up_proj_weight_quantizers.198 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.gate_up_proj_weight_quantizers.199 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.27e-01 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.gate_up_proj_weight_quantizers.200 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.gate_up_proj_weight_quantizers.201 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.17e-01 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.gate_up_proj_weight_quantizers.202 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.23e-01 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.gate_up_proj_weight_quantizers.203 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.34e-01 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.gate_up_proj_weight_quantizers.204 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.gate_up_proj_weight_quantizers.205 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.gate_up_proj_weight_quantizers.206 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.81e-01 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.gate_up_proj_weight_quantizers.207 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.96e-01 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.gate_up_proj_weight_quantizers.208 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.36e-01 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.gate_up_proj_weight_quantizers.209 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.43e-01 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.gate_up_proj_weight_quantizers.210 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.gate_up_proj_weight_quantizers.211 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.gate_up_proj_weight_quantizers.212 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.20e-01 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.gate_up_proj_weight_quantizers.213 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.17e-01 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.gate_up_proj_weight_quantizers.214 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.gate_up_proj_weight_quantizers.215 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.gate_up_proj_weight_quantizers.216 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.gate_up_proj_weight_quantizers.217 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.gate_up_proj_weight_quantizers.218 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.02e-01 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.gate_up_proj_weight_quantizers.219 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.54e-01 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.gate_up_proj_weight_quantizers.220 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.gate_up_proj_weight_quantizers.221 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.gate_up_proj_weight_quantizers.222 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.43e-01 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.gate_up_proj_weight_quantizers.223 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.88e-01 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.gate_up_proj_weight_quantizers.224 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.46e-01 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.gate_up_proj_weight_quantizers.225 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.gate_up_proj_weight_quantizers.226 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.24e-01 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.gate_up_proj_weight_quantizers.227 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.47e-01 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.gate_up_proj_weight_quantizers.228 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.gate_up_proj_weight_quantizers.229 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.gate_up_proj_weight_quantizers.230 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.83e-01 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.gate_up_proj_weight_quantizers.231 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.84e-01 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.gate_up_proj_weight_quantizers.232 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.gate_up_proj_weight_quantizers.233 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.17e-01 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.gate_up_proj_weight_quantizers.234 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.gate_up_proj_weight_quantizers.235 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.41e-01 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.gate_up_proj_weight_quantizers.236 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.14e-01 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.gate_up_proj_weight_quantizers.237 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.20e-01 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.gate_up_proj_weight_quantizers.238 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.gate_up_proj_weight_quantizers.239 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.05e-01 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.gate_up_proj_weight_quantizers.240 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.gate_up_proj_weight_quantizers.241 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.gate_up_proj_weight_quantizers.242 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.19e-01 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.gate_up_proj_weight_quantizers.243 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.gate_up_proj_weight_quantizers.244 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.64e-02 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.gate_up_proj_weight_quantizers.245 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.29e-01 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.gate_up_proj_weight_quantizers.246 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.gate_up_proj_weight_quantizers.247 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.41e-01 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.gate_up_proj_weight_quantizers.248 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.25e-01 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.gate_up_proj_weight_quantizers.249 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.gate_up_proj_weight_quantizers.250 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.18e-01 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.gate_up_proj_weight_quantizers.251 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.62e-01 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.gate_up_proj_weight_quantizers.252 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.gate_up_proj_weight_quantizers.253 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.gate_up_proj_weight_quantizers.254 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.44e-01 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.gate_up_proj_weight_quantizers.255 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.51e-01 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.down_proj_input_quantizer TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.52e+01 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.down_proj_weight_quantizers.0 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.59e-02 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.down_proj_weight_quantizers.1 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.down_proj_weight_quantizers.2 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.30e-02 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.down_proj_weight_quantizers.3 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.down_proj_weight_quantizers.4 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.down_proj_weight_quantizers.5 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.down_proj_weight_quantizers.6 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.40e-02 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.down_proj_weight_quantizers.7 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.down_proj_weight_quantizers.8 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.down_proj_weight_quantizers.9 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.64e-02 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.down_proj_weight_quantizers.10 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.59e-02 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.down_proj_weight_quantizers.11 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.45e-02 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.down_proj_weight_quantizers.12 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.down_proj_weight_quantizers.13 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.down_proj_weight_quantizers.14 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.down_proj_weight_quantizers.15 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.down_proj_weight_quantizers.16 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.down_proj_weight_quantizers.17 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.down_proj_weight_quantizers.18 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.down_proj_weight_quantizers.19 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.45e-02 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.down_proj_weight_quantizers.20 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.down_proj_weight_quantizers.21 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.down_proj_weight_quantizers.22 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.down_proj_weight_quantizers.23 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.down_proj_weight_quantizers.24 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.down_proj_weight_quantizers.25 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.down_proj_weight_quantizers.26 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.45e-02 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.down_proj_weight_quantizers.27 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.down_proj_weight_quantizers.28 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.21e-01 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.down_proj_weight_quantizers.29 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.29e-01 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.down_proj_weight_quantizers.30 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.31e-01 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.down_proj_weight_quantizers.31 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.down_proj_weight_quantizers.32 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.down_proj_weight_quantizers.33 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.down_proj_weight_quantizers.34 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.down_proj_weight_quantizers.35 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.down_proj_weight_quantizers.36 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.79e-01 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.down_proj_weight_quantizers.37 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.13e-01 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.down_proj_weight_quantizers.38 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.33e-01 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.down_proj_weight_quantizers.39 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.down_proj_weight_quantizers.40 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.down_proj_weight_quantizers.41 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.41e-01 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.down_proj_weight_quantizers.42 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.down_proj_weight_quantizers.43 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.54e-02 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.down_proj_weight_quantizers.44 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.down_proj_weight_quantizers.45 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.54e-02 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.down_proj_weight_quantizers.46 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.down_proj_weight_quantizers.47 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.down_proj_weight_quantizers.48 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.50e-02 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.down_proj_weight_quantizers.49 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.40e-02 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.down_proj_weight_quantizers.50 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.down_proj_weight_quantizers.51 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.28e-01 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.down_proj_weight_quantizers.52 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.down_proj_weight_quantizers.53 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.down_proj_weight_quantizers.54 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.down_proj_weight_quantizers.55 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.down_proj_weight_quantizers.56 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.13e-01 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.down_proj_weight_quantizers.57 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.down_proj_weight_quantizers.58 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.92e-01 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.down_proj_weight_quantizers.59 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.down_proj_weight_quantizers.60 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.down_proj_weight_quantizers.61 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.down_proj_weight_quantizers.62 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.down_proj_weight_quantizers.63 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.down_proj_weight_quantizers.64 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.down_proj_weight_quantizers.65 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.down_proj_weight_quantizers.66 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.down_proj_weight_quantizers.67 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.down_proj_weight_quantizers.68 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.down_proj_weight_quantizers.69 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.23e-01 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.down_proj_weight_quantizers.70 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.down_proj_weight_quantizers.71 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.down_proj_weight_quantizers.72 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.down_proj_weight_quantizers.73 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.down_proj_weight_quantizers.74 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.down_proj_weight_quantizers.75 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.17e-01 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.down_proj_weight_quantizers.76 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.down_proj_weight_quantizers.77 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.59e-02 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.down_proj_weight_quantizers.78 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.down_proj_weight_quantizers.79 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.down_proj_weight_quantizers.80 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=5.04e-01 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.down_proj_weight_quantizers.81 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.down_proj_weight_quantizers.82 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.00e-01 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.down_proj_weight_quantizers.83 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.50e-02 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.down_proj_weight_quantizers.84 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.down_proj_weight_quantizers.85 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.down_proj_weight_quantizers.86 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.down_proj_weight_quantizers.87 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.down_proj_weight_quantizers.88 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.71e-01 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.down_proj_weight_quantizers.89 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.down_proj_weight_quantizers.90 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.down_proj_weight_quantizers.91 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.down_proj_weight_quantizers.92 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.down_proj_weight_quantizers.93 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.down_proj_weight_quantizers.94 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.down_proj_weight_quantizers.95 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.down_proj_weight_quantizers.96 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.88e-01 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.down_proj_weight_quantizers.97 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.down_proj_weight_quantizers.98 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.down_proj_weight_quantizers.99 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.down_proj_weight_quantizers.100 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.down_proj_weight_quantizers.101 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.61e-01 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.down_proj_weight_quantizers.102 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.down_proj_weight_quantizers.103 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.down_proj_weight_quantizers.104 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.down_proj_weight_quantizers.105 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.down_proj_weight_quantizers.106 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.59e-02 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.down_proj_weight_quantizers.107 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.down_proj_weight_quantizers.108 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.down_proj_weight_quantizers.109 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.57e-01 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.down_proj_weight_quantizers.110 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.45e-01 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.down_proj_weight_quantizers.111 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.down_proj_weight_quantizers.112 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.down_proj_weight_quantizers.113 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.down_proj_weight_quantizers.114 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.down_proj_weight_quantizers.115 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.99e-01 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.down_proj_weight_quantizers.116 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.down_proj_weight_quantizers.117 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.down_proj_weight_quantizers.118 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.57e-01 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.down_proj_weight_quantizers.119 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.down_proj_weight_quantizers.120 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.down_proj_weight_quantizers.121 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.17e-01 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.down_proj_weight_quantizers.122 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.down_proj_weight_quantizers.123 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.54e-02 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.down_proj_weight_quantizers.124 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.down_proj_weight_quantizers.125 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.down_proj_weight_quantizers.126 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.39e-01 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.down_proj_weight_quantizers.127 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.down_proj_weight_quantizers.128 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.59e-02 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.down_proj_weight_quantizers.129 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.down_proj_weight_quantizers.130 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.down_proj_weight_quantizers.131 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.down_proj_weight_quantizers.132 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.11e-01 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.down_proj_weight_quantizers.133 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.26e-01 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.down_proj_weight_quantizers.134 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.down_proj_weight_quantizers.135 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.down_proj_weight_quantizers.136 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.85e-01 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.down_proj_weight_quantizers.137 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.down_proj_weight_quantizers.138 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.down_proj_weight_quantizers.139 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.79e-01 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.down_proj_weight_quantizers.140 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.down_proj_weight_quantizers.141 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.down_proj_weight_quantizers.142 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.down_proj_weight_quantizers.143 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.26e-01 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.down_proj_weight_quantizers.144 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.18e-01 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.down_proj_weight_quantizers.145 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.46e-01 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.down_proj_weight_quantizers.146 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.down_proj_weight_quantizers.147 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.down_proj_weight_quantizers.148 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.down_proj_weight_quantizers.149 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.down_proj_weight_quantizers.150 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.down_proj_weight_quantizers.151 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.38e-01 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.down_proj_weight_quantizers.152 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.down_proj_weight_quantizers.153 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.29e-01 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.down_proj_weight_quantizers.154 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.down_proj_weight_quantizers.155 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.26e-01 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.down_proj_weight_quantizers.156 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.06e-01 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.down_proj_weight_quantizers.157 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.down_proj_weight_quantizers.158 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.down_proj_weight_quantizers.159 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.31e-01 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.down_proj_weight_quantizers.160 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.down_proj_weight_quantizers.161 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.06e-01 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.down_proj_weight_quantizers.162 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.34e-01 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.down_proj_weight_quantizers.163 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.down_proj_weight_quantizers.164 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.06e-01 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.down_proj_weight_quantizers.165 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.down_proj_weight_quantizers.166 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.down_proj_weight_quantizers.167 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.down_proj_weight_quantizers.168 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.down_proj_weight_quantizers.169 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.down_proj_weight_quantizers.170 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.down_proj_weight_quantizers.171 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.down_proj_weight_quantizers.172 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.down_proj_weight_quantizers.173 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.down_proj_weight_quantizers.174 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.down_proj_weight_quantizers.175 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.down_proj_weight_quantizers.176 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.16e-01 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.down_proj_weight_quantizers.177 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.down_proj_weight_quantizers.178 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.down_proj_weight_quantizers.179 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.down_proj_weight_quantizers.180 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.down_proj_weight_quantizers.181 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.down_proj_weight_quantizers.182 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.67e-01 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.down_proj_weight_quantizers.183 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.down_proj_weight_quantizers.184 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.35e-02 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.down_proj_weight_quantizers.185 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.down_proj_weight_quantizers.186 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.down_proj_weight_quantizers.187 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.down_proj_weight_quantizers.188 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.down_proj_weight_quantizers.189 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.43e-01 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.down_proj_weight_quantizers.190 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.down_proj_weight_quantizers.191 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.down_proj_weight_quantizers.192 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.down_proj_weight_quantizers.193 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.40e-01 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.down_proj_weight_quantizers.194 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.60e-01 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.down_proj_weight_quantizers.195 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.down_proj_weight_quantizers.196 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.down_proj_weight_quantizers.197 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.down_proj_weight_quantizers.198 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.down_proj_weight_quantizers.199 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.60e-01 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.down_proj_weight_quantizers.200 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.down_proj_weight_quantizers.201 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.down_proj_weight_quantizers.202 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.25e-01 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.down_proj_weight_quantizers.203 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.down_proj_weight_quantizers.204 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.down_proj_weight_quantizers.205 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.down_proj_weight_quantizers.206 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.26e-01 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.down_proj_weight_quantizers.207 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.15e-01 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.down_proj_weight_quantizers.208 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.64e-02 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.down_proj_weight_quantizers.209 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.37e-01 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.down_proj_weight_quantizers.210 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.down_proj_weight_quantizers.211 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.50e-02 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.down_proj_weight_quantizers.212 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.down_proj_weight_quantizers.213 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.09e-01 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.down_proj_weight_quantizers.214 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.down_proj_weight_quantizers.215 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.50e-02 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.down_proj_weight_quantizers.216 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.down_proj_weight_quantizers.217 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.down_proj_weight_quantizers.218 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.81e-01 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.down_proj_weight_quantizers.219 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.88e-01 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.down_proj_weight_quantizers.220 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.down_proj_weight_quantizers.221 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.down_proj_weight_quantizers.222 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.14e-01 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.down_proj_weight_quantizers.223 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.17e-01 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.down_proj_weight_quantizers.224 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.down_proj_weight_quantizers.225 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.20e-02 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.down_proj_weight_quantizers.226 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.down_proj_weight_quantizers.227 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.39e-01 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.down_proj_weight_quantizers.228 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.down_proj_weight_quantizers.229 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.down_proj_weight_quantizers.230 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.down_proj_weight_quantizers.231 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.40e-01 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.down_proj_weight_quantizers.232 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.down_proj_weight_quantizers.233 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.32e-01 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.down_proj_weight_quantizers.234 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.54e-02 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.down_proj_weight_quantizers.235 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.44e-01 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.down_proj_weight_quantizers.236 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.35e-01 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.down_proj_weight_quantizers.237 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.08e-01 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.down_proj_weight_quantizers.238 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.64e-02 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.down_proj_weight_quantizers.239 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.88e-01 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.down_proj_weight_quantizers.240 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.down_proj_weight_quantizers.241 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.down_proj_weight_quantizers.242 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.down_proj_weight_quantizers.243 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.down_proj_weight_quantizers.244 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.down_proj_weight_quantizers.245 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.down_proj_weight_quantizers.246 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.20e-02 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.down_proj_weight_quantizers.247 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.down_proj_weight_quantizers.248 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.down_proj_weight_quantizers.249 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.50e-02 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.down_proj_weight_quantizers.250 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.98e-01 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.down_proj_weight_quantizers.251 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.down_proj_weight_quantizers.252 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.27e-01 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.down_proj_weight_quantizers.253 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.down_proj_weight_quantizers.254 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.down_proj_weight_quantizers.255 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.47.mlp.shared_experts.gate_proj.input_quantizer TensorQuantizer(disabled) +model.layers.47.mlp.shared_experts.gate_proj.output_quantizer TensorQuantizer(disabled) +model.layers.47.mlp.shared_experts.gate_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.47.mlp.shared_experts.up_proj.input_quantizer TensorQuantizer(disabled) +model.layers.47.mlp.shared_experts.up_proj.output_quantizer TensorQuantizer(disabled) +model.layers.47.mlp.shared_experts.up_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.47.mlp.shared_experts.down_proj.input_quantizer TensorQuantizer(disabled) +model.layers.47.mlp.shared_experts.down_proj.output_quantizer TensorQuantizer(disabled) +model.layers.47.mlp.shared_experts.down_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.48.self_attn.q_a_proj.input_quantizer TensorQuantizer(disabled) +model.layers.48.self_attn.q_a_proj.output_quantizer TensorQuantizer(disabled) +model.layers.48.self_attn.q_a_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.48.self_attn.q_b_proj.input_quantizer TensorQuantizer(disabled) +model.layers.48.self_attn.q_b_proj.output_quantizer TensorQuantizer(disabled) +model.layers.48.self_attn.q_b_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.48.self_attn.kv_a_proj_with_mqa.input_quantizer TensorQuantizer(disabled) +model.layers.48.self_attn.kv_a_proj_with_mqa.output_quantizer TensorQuantizer(disabled) +model.layers.48.self_attn.kv_a_proj_with_mqa.weight_quantizer TensorQuantizer(disabled) +model.layers.48.self_attn.kv_b_proj.input_quantizer TensorQuantizer(disabled) +model.layers.48.self_attn.kv_b_proj.output_quantizer TensorQuantizer(disabled) +model.layers.48.self_attn.kv_b_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.48.self_attn.o_proj.input_quantizer TensorQuantizer(disabled) +model.layers.48.self_attn.o_proj.output_quantizer TensorQuantizer(disabled) +model.layers.48.self_attn.o_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.48.self_attn.q_bmm_quantizer TensorQuantizer(disabled) +model.layers.48.self_attn.k_bmm_quantizer TensorQuantizer((4, 3) bit fake per-tensor amax=4.48e+02(const) calibrator=MaxCalibrator quant) +model.layers.48.self_attn.v_bmm_quantizer TensorQuantizer((4, 3) bit fake per-tensor amax=4.48e+02(const) calibrator=MaxCalibrator quant) +model.layers.48.self_attn.p_bmm_quantizer TensorQuantizer(disabled) +model.layers.48.mlp.experts.gate_up_proj_input_quantizer TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.30e+00 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.gate_up_proj_weight_quantizers.0 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.gate_up_proj_weight_quantizers.1 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.35e-01 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.gate_up_proj_weight_quantizers.2 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.gate_up_proj_weight_quantizers.3 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.62e-01 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.gate_up_proj_weight_quantizers.4 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.gate_up_proj_weight_quantizers.5 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.gate_up_proj_weight_quantizers.6 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.gate_up_proj_weight_quantizers.7 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.13e-01 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.gate_up_proj_weight_quantizers.8 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.46e-01 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.gate_up_proj_weight_quantizers.9 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.gate_up_proj_weight_quantizers.10 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.gate_up_proj_weight_quantizers.11 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.37e-01 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.gate_up_proj_weight_quantizers.12 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.gate_up_proj_weight_quantizers.13 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.30e-01 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.gate_up_proj_weight_quantizers.14 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.25e-01 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.gate_up_proj_weight_quantizers.15 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.gate_up_proj_weight_quantizers.16 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.57e-01 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.gate_up_proj_weight_quantizers.17 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.gate_up_proj_weight_quantizers.18 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.24e-01 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.gate_up_proj_weight_quantizers.19 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.gate_up_proj_weight_quantizers.20 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.74e-01 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.gate_up_proj_weight_quantizers.21 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.20e-01 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.gate_up_proj_weight_quantizers.22 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.30e-01 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.gate_up_proj_weight_quantizers.23 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.gate_up_proj_weight_quantizers.24 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.62e-01 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.gate_up_proj_weight_quantizers.25 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.gate_up_proj_weight_quantizers.26 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.16e-01 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.gate_up_proj_weight_quantizers.27 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.gate_up_proj_weight_quantizers.28 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.17e-01 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.gate_up_proj_weight_quantizers.29 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.gate_up_proj_weight_quantizers.30 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.gate_up_proj_weight_quantizers.31 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.gate_up_proj_weight_quantizers.32 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.31e-01 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.gate_up_proj_weight_quantizers.33 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.03e-01 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.gate_up_proj_weight_quantizers.34 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.gate_up_proj_weight_quantizers.35 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.gate_up_proj_weight_quantizers.36 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.19e-01 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.gate_up_proj_weight_quantizers.37 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.53e-01 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.gate_up_proj_weight_quantizers.38 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.gate_up_proj_weight_quantizers.39 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.66e-01 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.gate_up_proj_weight_quantizers.40 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.gate_up_proj_weight_quantizers.41 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.25e-01 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.gate_up_proj_weight_quantizers.42 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.64e-01 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.gate_up_proj_weight_quantizers.43 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.gate_up_proj_weight_quantizers.44 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.gate_up_proj_weight_quantizers.45 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.16e-01 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.gate_up_proj_weight_quantizers.46 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.13e-01 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.gate_up_proj_weight_quantizers.47 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.gate_up_proj_weight_quantizers.48 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.54e-01 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.gate_up_proj_weight_quantizers.49 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.gate_up_proj_weight_quantizers.50 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.15e-01 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.gate_up_proj_weight_quantizers.51 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.22e-01 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.gate_up_proj_weight_quantizers.52 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.41e-01 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.gate_up_proj_weight_quantizers.53 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.gate_up_proj_weight_quantizers.54 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.14e-01 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.gate_up_proj_weight_quantizers.55 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.gate_up_proj_weight_quantizers.56 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.65e-01 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.gate_up_proj_weight_quantizers.57 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.90e-01 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.gate_up_proj_weight_quantizers.58 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.39e-01 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.gate_up_proj_weight_quantizers.59 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.gate_up_proj_weight_quantizers.60 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.gate_up_proj_weight_quantizers.61 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.gate_up_proj_weight_quantizers.62 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.27e-01 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.gate_up_proj_weight_quantizers.63 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.gate_up_proj_weight_quantizers.64 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.28e-01 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.gate_up_proj_weight_quantizers.65 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.13e-01 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.gate_up_proj_weight_quantizers.66 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.gate_up_proj_weight_quantizers.67 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.gate_up_proj_weight_quantizers.68 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.gate_up_proj_weight_quantizers.69 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.31e-01 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.gate_up_proj_weight_quantizers.70 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.gate_up_proj_weight_quantizers.71 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.gate_up_proj_weight_quantizers.72 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.gate_up_proj_weight_quantizers.73 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.09e-01 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.gate_up_proj_weight_quantizers.74 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.24e-01 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.gate_up_proj_weight_quantizers.75 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.65e-01 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.gate_up_proj_weight_quantizers.76 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.gate_up_proj_weight_quantizers.77 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.95e-01 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.gate_up_proj_weight_quantizers.78 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.37e-01 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.gate_up_proj_weight_quantizers.79 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.gate_up_proj_weight_quantizers.80 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.gate_up_proj_weight_quantizers.81 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.gate_up_proj_weight_quantizers.82 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.gate_up_proj_weight_quantizers.83 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.gate_up_proj_weight_quantizers.84 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.gate_up_proj_weight_quantizers.85 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.88e-01 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.gate_up_proj_weight_quantizers.86 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.gate_up_proj_weight_quantizers.87 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.gate_up_proj_weight_quantizers.88 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.gate_up_proj_weight_quantizers.89 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.gate_up_proj_weight_quantizers.90 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.gate_up_proj_weight_quantizers.91 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.gate_up_proj_weight_quantizers.92 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.29e-01 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.gate_up_proj_weight_quantizers.93 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.32e-01 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.gate_up_proj_weight_quantizers.94 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.95e-01 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.gate_up_proj_weight_quantizers.95 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.gate_up_proj_weight_quantizers.96 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.32e-01 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.gate_up_proj_weight_quantizers.97 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.21e-01 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.gate_up_proj_weight_quantizers.98 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.gate_up_proj_weight_quantizers.99 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.gate_up_proj_weight_quantizers.100 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.70e-01 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.gate_up_proj_weight_quantizers.101 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.26e-01 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.gate_up_proj_weight_quantizers.102 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.gate_up_proj_weight_quantizers.103 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.gate_up_proj_weight_quantizers.104 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.18e-01 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.gate_up_proj_weight_quantizers.105 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.28e-01 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.gate_up_proj_weight_quantizers.106 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.gate_up_proj_weight_quantizers.107 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.19e-01 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.gate_up_proj_weight_quantizers.108 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.gate_up_proj_weight_quantizers.109 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.gate_up_proj_weight_quantizers.110 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.gate_up_proj_weight_quantizers.111 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.14e-01 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.gate_up_proj_weight_quantizers.112 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.gate_up_proj_weight_quantizers.113 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.gate_up_proj_weight_quantizers.114 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.24e-01 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.gate_up_proj_weight_quantizers.115 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.52e-01 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.gate_up_proj_weight_quantizers.116 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.24e-01 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.gate_up_proj_weight_quantizers.117 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.gate_up_proj_weight_quantizers.118 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.gate_up_proj_weight_quantizers.119 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.16e-01 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.gate_up_proj_weight_quantizers.120 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.14e-01 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.gate_up_proj_weight_quantizers.121 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.43e-01 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.gate_up_proj_weight_quantizers.122 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.47e-01 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.gate_up_proj_weight_quantizers.123 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.gate_up_proj_weight_quantizers.124 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.19e-01 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.gate_up_proj_weight_quantizers.125 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.gate_up_proj_weight_quantizers.126 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.47e-01 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.gate_up_proj_weight_quantizers.127 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.39e-01 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.gate_up_proj_weight_quantizers.128 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=4.82e-01 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.gate_up_proj_weight_quantizers.129 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.19e-01 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.gate_up_proj_weight_quantizers.130 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.gate_up_proj_weight_quantizers.131 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.gate_up_proj_weight_quantizers.132 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.gate_up_proj_weight_quantizers.133 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.gate_up_proj_weight_quantizers.134 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.gate_up_proj_weight_quantizers.135 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.gate_up_proj_weight_quantizers.136 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.gate_up_proj_weight_quantizers.137 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.gate_up_proj_weight_quantizers.138 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.gate_up_proj_weight_quantizers.139 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.gate_up_proj_weight_quantizers.140 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.gate_up_proj_weight_quantizers.141 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.gate_up_proj_weight_quantizers.142 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.gate_up_proj_weight_quantizers.143 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.13e-01 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.gate_up_proj_weight_quantizers.144 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.22e-01 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.gate_up_proj_weight_quantizers.145 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.25e-01 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.gate_up_proj_weight_quantizers.146 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.gate_up_proj_weight_quantizers.147 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.gate_up_proj_weight_quantizers.148 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.08e-01 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.gate_up_proj_weight_quantizers.149 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.gate_up_proj_weight_quantizers.150 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.16e-01 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.gate_up_proj_weight_quantizers.151 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.52e-01 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.gate_up_proj_weight_quantizers.152 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.13e-01 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.gate_up_proj_weight_quantizers.153 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.17e-01 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.gate_up_proj_weight_quantizers.154 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.33e-01 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.gate_up_proj_weight_quantizers.155 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.gate_up_proj_weight_quantizers.156 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.gate_up_proj_weight_quantizers.157 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.14e-01 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.gate_up_proj_weight_quantizers.158 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.gate_up_proj_weight_quantizers.159 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.gate_up_proj_weight_quantizers.160 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.gate_up_proj_weight_quantizers.161 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.gate_up_proj_weight_quantizers.162 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.14e-01 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.gate_up_proj_weight_quantizers.163 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.gate_up_proj_weight_quantizers.164 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.40e-01 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.gate_up_proj_weight_quantizers.165 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.gate_up_proj_weight_quantizers.166 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.gate_up_proj_weight_quantizers.167 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.gate_up_proj_weight_quantizers.168 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.gate_up_proj_weight_quantizers.169 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.gate_up_proj_weight_quantizers.170 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.91e-01 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.gate_up_proj_weight_quantizers.171 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.gate_up_proj_weight_quantizers.172 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.17e-01 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.gate_up_proj_weight_quantizers.173 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.gate_up_proj_weight_quantizers.174 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.gate_up_proj_weight_quantizers.175 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.25e-01 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.gate_up_proj_weight_quantizers.176 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.34e-01 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.gate_up_proj_weight_quantizers.177 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.19e-01 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.gate_up_proj_weight_quantizers.178 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.gate_up_proj_weight_quantizers.179 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.69e-01 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.gate_up_proj_weight_quantizers.180 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.gate_up_proj_weight_quantizers.181 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.18e-01 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.gate_up_proj_weight_quantizers.182 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.gate_up_proj_weight_quantizers.183 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.gate_up_proj_weight_quantizers.184 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.gate_up_proj_weight_quantizers.185 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.91e-01 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.gate_up_proj_weight_quantizers.186 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.34e-01 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.gate_up_proj_weight_quantizers.187 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.68e-01 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.gate_up_proj_weight_quantizers.188 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.15e-01 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.gate_up_proj_weight_quantizers.189 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.gate_up_proj_weight_quantizers.190 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.17e-01 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.gate_up_proj_weight_quantizers.191 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.14e-01 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.gate_up_proj_weight_quantizers.192 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.16e-01 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.gate_up_proj_weight_quantizers.193 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.gate_up_proj_weight_quantizers.194 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.33e-01 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.gate_up_proj_weight_quantizers.195 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.27e-01 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.gate_up_proj_weight_quantizers.196 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.gate_up_proj_weight_quantizers.197 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.32e-01 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.gate_up_proj_weight_quantizers.198 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.gate_up_proj_weight_quantizers.199 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.13e-01 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.gate_up_proj_weight_quantizers.200 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.gate_up_proj_weight_quantizers.201 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.gate_up_proj_weight_quantizers.202 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.gate_up_proj_weight_quantizers.203 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.21e-01 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.gate_up_proj_weight_quantizers.204 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.gate_up_proj_weight_quantizers.205 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.gate_up_proj_weight_quantizers.206 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.gate_up_proj_weight_quantizers.207 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.gate_up_proj_weight_quantizers.208 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.46e-01 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.gate_up_proj_weight_quantizers.209 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.17e-01 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.gate_up_proj_weight_quantizers.210 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.23e-01 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.gate_up_proj_weight_quantizers.211 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.25e-01 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.gate_up_proj_weight_quantizers.212 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.gate_up_proj_weight_quantizers.213 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.31e-01 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.gate_up_proj_weight_quantizers.214 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.gate_up_proj_weight_quantizers.215 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.65e-01 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.gate_up_proj_weight_quantizers.216 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.gate_up_proj_weight_quantizers.217 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.94e-01 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.gate_up_proj_weight_quantizers.218 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.36e-01 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.gate_up_proj_weight_quantizers.219 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.gate_up_proj_weight_quantizers.220 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.gate_up_proj_weight_quantizers.221 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.gate_up_proj_weight_quantizers.222 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.54e-01 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.gate_up_proj_weight_quantizers.223 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.gate_up_proj_weight_quantizers.224 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.60e-01 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.gate_up_proj_weight_quantizers.225 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.gate_up_proj_weight_quantizers.226 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.18e-01 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.gate_up_proj_weight_quantizers.227 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.20e-01 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.gate_up_proj_weight_quantizers.228 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.13e-01 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.gate_up_proj_weight_quantizers.229 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.15e-01 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.gate_up_proj_weight_quantizers.230 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.97e-01 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.gate_up_proj_weight_quantizers.231 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.gate_up_proj_weight_quantizers.232 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.50e-01 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.gate_up_proj_weight_quantizers.233 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.67e-01 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.gate_up_proj_weight_quantizers.234 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.22e-01 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.gate_up_proj_weight_quantizers.235 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.gate_up_proj_weight_quantizers.236 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.17e-01 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.gate_up_proj_weight_quantizers.237 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.gate_up_proj_weight_quantizers.238 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.gate_up_proj_weight_quantizers.239 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.58e-01 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.gate_up_proj_weight_quantizers.240 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.43e-01 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.gate_up_proj_weight_quantizers.241 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.gate_up_proj_weight_quantizers.242 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.13e-01 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.gate_up_proj_weight_quantizers.243 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.gate_up_proj_weight_quantizers.244 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.gate_up_proj_weight_quantizers.245 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.30e-01 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.gate_up_proj_weight_quantizers.246 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.gate_up_proj_weight_quantizers.247 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.32e-01 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.gate_up_proj_weight_quantizers.248 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.86e-01 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.gate_up_proj_weight_quantizers.249 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.gate_up_proj_weight_quantizers.250 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.gate_up_proj_weight_quantizers.251 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.75e-01 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.gate_up_proj_weight_quantizers.252 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.32e-01 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.gate_up_proj_weight_quantizers.253 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.gate_up_proj_weight_quantizers.254 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.15e-01 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.gate_up_proj_weight_quantizers.255 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.74e-01 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.down_proj_input_quantizer TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.75e+01 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.down_proj_weight_quantizers.0 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.down_proj_weight_quantizers.1 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.59e-02 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.down_proj_weight_quantizers.2 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.down_proj_weight_quantizers.3 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.down_proj_weight_quantizers.4 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.down_proj_weight_quantizers.5 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.down_proj_weight_quantizers.6 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.down_proj_weight_quantizers.7 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.down_proj_weight_quantizers.8 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.down_proj_weight_quantizers.9 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.down_proj_weight_quantizers.10 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.59e-02 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.down_proj_weight_quantizers.11 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.94e-01 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.down_proj_weight_quantizers.12 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.down_proj_weight_quantizers.13 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.down_proj_weight_quantizers.14 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.55e-01 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.down_proj_weight_quantizers.15 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.down_proj_weight_quantizers.16 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.down_proj_weight_quantizers.17 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.down_proj_weight_quantizers.18 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.down_proj_weight_quantizers.19 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.down_proj_weight_quantizers.20 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.14e-01 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.down_proj_weight_quantizers.21 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.down_proj_weight_quantizers.22 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.30e-01 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.down_proj_weight_quantizers.23 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.45e-02 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.down_proj_weight_quantizers.24 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.down_proj_weight_quantizers.25 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.down_proj_weight_quantizers.26 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.down_proj_weight_quantizers.27 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.down_proj_weight_quantizers.28 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.down_proj_weight_quantizers.29 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.64e-02 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.down_proj_weight_quantizers.30 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.down_proj_weight_quantizers.31 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.down_proj_weight_quantizers.32 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.down_proj_weight_quantizers.33 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.02e-01 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.down_proj_weight_quantizers.34 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.down_proj_weight_quantizers.35 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.down_proj_weight_quantizers.36 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.down_proj_weight_quantizers.37 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.21e-01 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.down_proj_weight_quantizers.38 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.down_proj_weight_quantizers.39 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.07e-01 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.down_proj_weight_quantizers.40 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.down_proj_weight_quantizers.41 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.down_proj_weight_quantizers.42 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.89e-01 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.down_proj_weight_quantizers.43 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.down_proj_weight_quantizers.44 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.down_proj_weight_quantizers.45 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.down_proj_weight_quantizers.46 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.down_proj_weight_quantizers.47 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.down_proj_weight_quantizers.48 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.22e-01 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.down_proj_weight_quantizers.49 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.down_proj_weight_quantizers.50 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.down_proj_weight_quantizers.51 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.down_proj_weight_quantizers.52 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=5.39e-01 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.down_proj_weight_quantizers.53 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.down_proj_weight_quantizers.54 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.down_proj_weight_quantizers.55 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.down_proj_weight_quantizers.56 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.80e-01 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.down_proj_weight_quantizers.57 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.down_proj_weight_quantizers.58 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.down_proj_weight_quantizers.59 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.40e-02 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.down_proj_weight_quantizers.60 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.64e-02 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.down_proj_weight_quantizers.61 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.down_proj_weight_quantizers.62 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.down_proj_weight_quantizers.63 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.down_proj_weight_quantizers.64 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.down_proj_weight_quantizers.65 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.down_proj_weight_quantizers.66 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.down_proj_weight_quantizers.67 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.23e-01 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.down_proj_weight_quantizers.68 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.42e-01 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.down_proj_weight_quantizers.69 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.52e-01 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.down_proj_weight_quantizers.70 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.down_proj_weight_quantizers.71 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.down_proj_weight_quantizers.72 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.51e-01 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.down_proj_weight_quantizers.73 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.38e-01 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.down_proj_weight_quantizers.74 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.down_proj_weight_quantizers.75 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.down_proj_weight_quantizers.76 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.50e-02 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.down_proj_weight_quantizers.77 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.83e-01 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.down_proj_weight_quantizers.78 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.27e-01 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.down_proj_weight_quantizers.79 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.down_proj_weight_quantizers.80 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.down_proj_weight_quantizers.81 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.down_proj_weight_quantizers.82 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.64e-02 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.down_proj_weight_quantizers.83 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.down_proj_weight_quantizers.84 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.down_proj_weight_quantizers.85 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.down_proj_weight_quantizers.86 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.16e-01 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.down_proj_weight_quantizers.87 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.down_proj_weight_quantizers.88 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.down_proj_weight_quantizers.89 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.down_proj_weight_quantizers.90 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.down_proj_weight_quantizers.91 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.down_proj_weight_quantizers.92 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.down_proj_weight_quantizers.93 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.down_proj_weight_quantizers.94 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.down_proj_weight_quantizers.95 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.down_proj_weight_quantizers.96 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.down_proj_weight_quantizers.97 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.down_proj_weight_quantizers.98 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.36e-01 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.down_proj_weight_quantizers.99 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.down_proj_weight_quantizers.100 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.30e-01 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.down_proj_weight_quantizers.101 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.54e-02 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.down_proj_weight_quantizers.102 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.down_proj_weight_quantizers.103 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.64e-02 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.down_proj_weight_quantizers.104 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.59e-02 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.down_proj_weight_quantizers.105 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.59e-02 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.down_proj_weight_quantizers.106 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.25e-01 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.down_proj_weight_quantizers.107 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.down_proj_weight_quantizers.108 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.36e-01 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.down_proj_weight_quantizers.109 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.down_proj_weight_quantizers.110 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.down_proj_weight_quantizers.111 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.down_proj_weight_quantizers.112 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.down_proj_weight_quantizers.113 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.down_proj_weight_quantizers.114 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.68e-01 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.down_proj_weight_quantizers.115 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.15e-01 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.down_proj_weight_quantizers.116 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.down_proj_weight_quantizers.117 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.down_proj_weight_quantizers.118 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.down_proj_weight_quantizers.119 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.down_proj_weight_quantizers.120 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.24e-01 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.down_proj_weight_quantizers.121 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.27e-01 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.down_proj_weight_quantizers.122 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.15e-01 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.down_proj_weight_quantizers.123 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.down_proj_weight_quantizers.124 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.down_proj_weight_quantizers.125 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.down_proj_weight_quantizers.126 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.13e-01 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.down_proj_weight_quantizers.127 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.34e-01 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.down_proj_weight_quantizers.128 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.50e-01 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.down_proj_weight_quantizers.129 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.down_proj_weight_quantizers.130 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.64e-02 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.down_proj_weight_quantizers.131 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.down_proj_weight_quantizers.132 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.down_proj_weight_quantizers.133 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.25e-01 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.down_proj_weight_quantizers.134 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.down_proj_weight_quantizers.135 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.down_proj_weight_quantizers.136 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.down_proj_weight_quantizers.137 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.down_proj_weight_quantizers.138 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.down_proj_weight_quantizers.139 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.54e-02 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.down_proj_weight_quantizers.140 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.down_proj_weight_quantizers.141 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.down_proj_weight_quantizers.142 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.down_proj_weight_quantizers.143 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.down_proj_weight_quantizers.144 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.down_proj_weight_quantizers.145 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.37e-01 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.down_proj_weight_quantizers.146 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.down_proj_weight_quantizers.147 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.down_proj_weight_quantizers.148 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.down_proj_weight_quantizers.149 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.down_proj_weight_quantizers.150 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.45e-02 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.down_proj_weight_quantizers.151 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.down_proj_weight_quantizers.152 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.down_proj_weight_quantizers.153 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.down_proj_weight_quantizers.154 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.down_proj_weight_quantizers.155 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.down_proj_weight_quantizers.156 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.down_proj_weight_quantizers.157 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.down_proj_weight_quantizers.158 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.down_proj_weight_quantizers.159 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.down_proj_weight_quantizers.160 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.down_proj_weight_quantizers.161 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.down_proj_weight_quantizers.162 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.down_proj_weight_quantizers.163 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.down_proj_weight_quantizers.164 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.78e-01 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.down_proj_weight_quantizers.165 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.down_proj_weight_quantizers.166 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.down_proj_weight_quantizers.167 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.88e-01 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.down_proj_weight_quantizers.168 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.down_proj_weight_quantizers.169 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.down_proj_weight_quantizers.170 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.97e-01 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.down_proj_weight_quantizers.171 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.down_proj_weight_quantizers.172 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.down_proj_weight_quantizers.173 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.down_proj_weight_quantizers.174 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.40e-02 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.down_proj_weight_quantizers.175 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.down_proj_weight_quantizers.176 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.down_proj_weight_quantizers.177 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.down_proj_weight_quantizers.178 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.59e-02 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.down_proj_weight_quantizers.179 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.03e-01 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.down_proj_weight_quantizers.180 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.24e-01 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.down_proj_weight_quantizers.181 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.down_proj_weight_quantizers.182 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.64e-02 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.down_proj_weight_quantizers.183 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.down_proj_weight_quantizers.184 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.down_proj_weight_quantizers.185 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.59e-02 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.down_proj_weight_quantizers.186 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.down_proj_weight_quantizers.187 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.41e-01 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.down_proj_weight_quantizers.188 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.down_proj_weight_quantizers.189 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.97e-01 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.down_proj_weight_quantizers.190 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.down_proj_weight_quantizers.191 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.down_proj_weight_quantizers.192 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.down_proj_weight_quantizers.193 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.50e-02 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.down_proj_weight_quantizers.194 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.24e-01 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.down_proj_weight_quantizers.195 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.64e-02 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.down_proj_weight_quantizers.196 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.down_proj_weight_quantizers.197 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.down_proj_weight_quantizers.198 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.64e-02 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.down_proj_weight_quantizers.199 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.down_proj_weight_quantizers.200 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.down_proj_weight_quantizers.201 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.down_proj_weight_quantizers.202 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.down_proj_weight_quantizers.203 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.down_proj_weight_quantizers.204 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.down_proj_weight_quantizers.205 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.down_proj_weight_quantizers.206 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.59e-02 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.down_proj_weight_quantizers.207 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.down_proj_weight_quantizers.208 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.down_proj_weight_quantizers.209 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.down_proj_weight_quantizers.210 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.down_proj_weight_quantizers.211 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.down_proj_weight_quantizers.212 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.down_proj_weight_quantizers.213 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.down_proj_weight_quantizers.214 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.50e-02 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.down_proj_weight_quantizers.215 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.down_proj_weight_quantizers.216 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.down_proj_weight_quantizers.217 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.68e-01 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.down_proj_weight_quantizers.218 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.down_proj_weight_quantizers.219 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.64e-02 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.down_proj_weight_quantizers.220 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.down_proj_weight_quantizers.221 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.down_proj_weight_quantizers.222 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.down_proj_weight_quantizers.223 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.down_proj_weight_quantizers.224 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.64e-02 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.down_proj_weight_quantizers.225 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.down_proj_weight_quantizers.226 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.down_proj_weight_quantizers.227 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.21e-01 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.down_proj_weight_quantizers.228 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.down_proj_weight_quantizers.229 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.down_proj_weight_quantizers.230 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.down_proj_weight_quantizers.231 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.down_proj_weight_quantizers.232 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.49e-01 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.down_proj_weight_quantizers.233 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.down_proj_weight_quantizers.234 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.down_proj_weight_quantizers.235 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.down_proj_weight_quantizers.236 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.down_proj_weight_quantizers.237 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.down_proj_weight_quantizers.238 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.54e-02 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.down_proj_weight_quantizers.239 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.down_proj_weight_quantizers.240 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.down_proj_weight_quantizers.241 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.down_proj_weight_quantizers.242 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.down_proj_weight_quantizers.243 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.down_proj_weight_quantizers.244 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.down_proj_weight_quantizers.245 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.down_proj_weight_quantizers.246 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.down_proj_weight_quantizers.247 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.down_proj_weight_quantizers.248 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.22e-01 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.down_proj_weight_quantizers.249 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.down_proj_weight_quantizers.250 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.down_proj_weight_quantizers.251 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.down_proj_weight_quantizers.252 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.down_proj_weight_quantizers.253 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.down_proj_weight_quantizers.254 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.down_proj_weight_quantizers.255 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.48.mlp.shared_experts.gate_proj.input_quantizer TensorQuantizer(disabled) +model.layers.48.mlp.shared_experts.gate_proj.output_quantizer TensorQuantizer(disabled) +model.layers.48.mlp.shared_experts.gate_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.48.mlp.shared_experts.up_proj.input_quantizer TensorQuantizer(disabled) +model.layers.48.mlp.shared_experts.up_proj.output_quantizer TensorQuantizer(disabled) +model.layers.48.mlp.shared_experts.up_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.48.mlp.shared_experts.down_proj.input_quantizer TensorQuantizer(disabled) +model.layers.48.mlp.shared_experts.down_proj.output_quantizer TensorQuantizer(disabled) +model.layers.48.mlp.shared_experts.down_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.49.self_attn.q_a_proj.input_quantizer TensorQuantizer(disabled) +model.layers.49.self_attn.q_a_proj.output_quantizer TensorQuantizer(disabled) +model.layers.49.self_attn.q_a_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.49.self_attn.q_b_proj.input_quantizer TensorQuantizer(disabled) +model.layers.49.self_attn.q_b_proj.output_quantizer TensorQuantizer(disabled) +model.layers.49.self_attn.q_b_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.49.self_attn.kv_a_proj_with_mqa.input_quantizer TensorQuantizer(disabled) +model.layers.49.self_attn.kv_a_proj_with_mqa.output_quantizer TensorQuantizer(disabled) +model.layers.49.self_attn.kv_a_proj_with_mqa.weight_quantizer TensorQuantizer(disabled) +model.layers.49.self_attn.kv_b_proj.input_quantizer TensorQuantizer(disabled) +model.layers.49.self_attn.kv_b_proj.output_quantizer TensorQuantizer(disabled) +model.layers.49.self_attn.kv_b_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.49.self_attn.o_proj.input_quantizer TensorQuantizer(disabled) +model.layers.49.self_attn.o_proj.output_quantizer TensorQuantizer(disabled) +model.layers.49.self_attn.o_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.49.self_attn.q_bmm_quantizer TensorQuantizer(disabled) +model.layers.49.self_attn.k_bmm_quantizer TensorQuantizer((4, 3) bit fake per-tensor amax=4.48e+02(const) calibrator=MaxCalibrator quant) +model.layers.49.self_attn.v_bmm_quantizer TensorQuantizer((4, 3) bit fake per-tensor amax=4.48e+02(const) calibrator=MaxCalibrator quant) +model.layers.49.self_attn.p_bmm_quantizer TensorQuantizer(disabled) +model.layers.49.mlp.experts.gate_up_proj_input_quantizer TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.59e+00 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.gate_up_proj_weight_quantizers.0 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.gate_up_proj_weight_quantizers.1 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.gate_up_proj_weight_quantizers.2 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.gate_up_proj_weight_quantizers.3 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.gate_up_proj_weight_quantizers.4 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.gate_up_proj_weight_quantizers.5 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.22e-01 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.gate_up_proj_weight_quantizers.6 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.gate_up_proj_weight_quantizers.7 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.25e-01 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.gate_up_proj_weight_quantizers.8 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.46e-01 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.gate_up_proj_weight_quantizers.9 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.gate_up_proj_weight_quantizers.10 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.17e-01 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.gate_up_proj_weight_quantizers.11 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.06e-01 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.gate_up_proj_weight_quantizers.12 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.01e-01 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.gate_up_proj_weight_quantizers.13 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.gate_up_proj_weight_quantizers.14 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.10e-01 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.gate_up_proj_weight_quantizers.15 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.gate_up_proj_weight_quantizers.16 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.17e-01 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.gate_up_proj_weight_quantizers.17 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.57e-01 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.gate_up_proj_weight_quantizers.18 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.46e-01 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.gate_up_proj_weight_quantizers.19 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.66e-01 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.gate_up_proj_weight_quantizers.20 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.32e-01 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.gate_up_proj_weight_quantizers.21 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.gate_up_proj_weight_quantizers.22 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.94e-01 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.gate_up_proj_weight_quantizers.23 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.81e-01 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.gate_up_proj_weight_quantizers.24 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.gate_up_proj_weight_quantizers.25 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.gate_up_proj_weight_quantizers.26 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.gate_up_proj_weight_quantizers.27 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.13e-01 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.gate_up_proj_weight_quantizers.28 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.gate_up_proj_weight_quantizers.29 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.gate_up_proj_weight_quantizers.30 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.gate_up_proj_weight_quantizers.31 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.gate_up_proj_weight_quantizers.32 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.gate_up_proj_weight_quantizers.33 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.23e-01 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.gate_up_proj_weight_quantizers.34 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.20e-01 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.gate_up_proj_weight_quantizers.35 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.gate_up_proj_weight_quantizers.36 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.gate_up_proj_weight_quantizers.37 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.gate_up_proj_weight_quantizers.38 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.23e-01 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.gate_up_proj_weight_quantizers.39 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.gate_up_proj_weight_quantizers.40 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.86e-01 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.gate_up_proj_weight_quantizers.41 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.gate_up_proj_weight_quantizers.42 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.gate_up_proj_weight_quantizers.43 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.28e-01 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.gate_up_proj_weight_quantizers.44 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.46e-01 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.gate_up_proj_weight_quantizers.45 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.32e-01 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.gate_up_proj_weight_quantizers.46 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.gate_up_proj_weight_quantizers.47 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.gate_up_proj_weight_quantizers.48 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.gate_up_proj_weight_quantizers.49 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.27e-01 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.gate_up_proj_weight_quantizers.50 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.16e-01 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.gate_up_proj_weight_quantizers.51 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.gate_up_proj_weight_quantizers.52 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.gate_up_proj_weight_quantizers.53 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.gate_up_proj_weight_quantizers.54 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.63e-01 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.gate_up_proj_weight_quantizers.55 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.gate_up_proj_weight_quantizers.56 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.gate_up_proj_weight_quantizers.57 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.gate_up_proj_weight_quantizers.58 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.64e-01 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.gate_up_proj_weight_quantizers.59 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.gate_up_proj_weight_quantizers.60 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.gate_up_proj_weight_quantizers.61 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.gate_up_proj_weight_quantizers.62 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.14e-01 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.gate_up_proj_weight_quantizers.63 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.62e-01 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.gate_up_proj_weight_quantizers.64 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.84e-01 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.gate_up_proj_weight_quantizers.65 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.gate_up_proj_weight_quantizers.66 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.gate_up_proj_weight_quantizers.67 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.19e-01 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.gate_up_proj_weight_quantizers.68 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.gate_up_proj_weight_quantizers.69 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.27e-01 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.gate_up_proj_weight_quantizers.70 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.gate_up_proj_weight_quantizers.71 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.gate_up_proj_weight_quantizers.72 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.54e-01 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.gate_up_proj_weight_quantizers.73 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.gate_up_proj_weight_quantizers.74 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.18e-01 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.gate_up_proj_weight_quantizers.75 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.gate_up_proj_weight_quantizers.76 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.17e-01 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.gate_up_proj_weight_quantizers.77 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.24e-01 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.gate_up_proj_weight_quantizers.78 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.gate_up_proj_weight_quantizers.79 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.gate_up_proj_weight_quantizers.80 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.gate_up_proj_weight_quantizers.81 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.gate_up_proj_weight_quantizers.82 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.gate_up_proj_weight_quantizers.83 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.29e-01 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.gate_up_proj_weight_quantizers.84 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.gate_up_proj_weight_quantizers.85 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.gate_up_proj_weight_quantizers.86 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.25e-01 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.gate_up_proj_weight_quantizers.87 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.24e-01 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.gate_up_proj_weight_quantizers.88 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.gate_up_proj_weight_quantizers.89 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.68e-01 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.gate_up_proj_weight_quantizers.90 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.gate_up_proj_weight_quantizers.91 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.gate_up_proj_weight_quantizers.92 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.gate_up_proj_weight_quantizers.93 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.gate_up_proj_weight_quantizers.94 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.gate_up_proj_weight_quantizers.95 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.29e-01 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.gate_up_proj_weight_quantizers.96 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.14e-01 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.gate_up_proj_weight_quantizers.97 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.27e-01 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.gate_up_proj_weight_quantizers.98 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.gate_up_proj_weight_quantizers.99 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.gate_up_proj_weight_quantizers.100 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.gate_up_proj_weight_quantizers.101 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.gate_up_proj_weight_quantizers.102 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.75e-01 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.gate_up_proj_weight_quantizers.103 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.14e-01 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.gate_up_proj_weight_quantizers.104 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.15e-01 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.gate_up_proj_weight_quantizers.105 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.58e-01 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.gate_up_proj_weight_quantizers.106 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.gate_up_proj_weight_quantizers.107 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.gate_up_proj_weight_quantizers.108 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.40e-01 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.gate_up_proj_weight_quantizers.109 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.33e-01 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.gate_up_proj_weight_quantizers.110 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.gate_up_proj_weight_quantizers.111 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.03e-01 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.gate_up_proj_weight_quantizers.112 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.15e-01 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.gate_up_proj_weight_quantizers.113 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.gate_up_proj_weight_quantizers.114 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.31e-01 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.gate_up_proj_weight_quantizers.115 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.gate_up_proj_weight_quantizers.116 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.03e-01 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.gate_up_proj_weight_quantizers.117 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.gate_up_proj_weight_quantizers.118 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.gate_up_proj_weight_quantizers.119 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.29e-01 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.gate_up_proj_weight_quantizers.120 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.gate_up_proj_weight_quantizers.121 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.gate_up_proj_weight_quantizers.122 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.54e-01 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.gate_up_proj_weight_quantizers.123 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.35e-01 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.gate_up_proj_weight_quantizers.124 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.gate_up_proj_weight_quantizers.125 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.gate_up_proj_weight_quantizers.126 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.71e-01 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.gate_up_proj_weight_quantizers.127 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.15e-01 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.gate_up_proj_weight_quantizers.128 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.13e-01 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.gate_up_proj_weight_quantizers.129 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.25e-01 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.gate_up_proj_weight_quantizers.130 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.32e-01 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.gate_up_proj_weight_quantizers.131 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.gate_up_proj_weight_quantizers.132 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.gate_up_proj_weight_quantizers.133 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.gate_up_proj_weight_quantizers.134 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.gate_up_proj_weight_quantizers.135 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.gate_up_proj_weight_quantizers.136 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.gate_up_proj_weight_quantizers.137 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.gate_up_proj_weight_quantizers.138 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.19e-01 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.gate_up_proj_weight_quantizers.139 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.14e-01 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.gate_up_proj_weight_quantizers.140 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.gate_up_proj_weight_quantizers.141 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.21e-01 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.gate_up_proj_weight_quantizers.142 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.gate_up_proj_weight_quantizers.143 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.gate_up_proj_weight_quantizers.144 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.14e-01 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.gate_up_proj_weight_quantizers.145 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.34e-01 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.gate_up_proj_weight_quantizers.146 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.37e-01 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.gate_up_proj_weight_quantizers.147 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.gate_up_proj_weight_quantizers.148 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.gate_up_proj_weight_quantizers.149 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.gate_up_proj_weight_quantizers.150 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.21e-01 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.gate_up_proj_weight_quantizers.151 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.39e-01 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.gate_up_proj_weight_quantizers.152 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.16e-01 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.gate_up_proj_weight_quantizers.153 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.gate_up_proj_weight_quantizers.154 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.gate_up_proj_weight_quantizers.155 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.gate_up_proj_weight_quantizers.156 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.gate_up_proj_weight_quantizers.157 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.gate_up_proj_weight_quantizers.158 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.gate_up_proj_weight_quantizers.159 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.64e-01 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.gate_up_proj_weight_quantizers.160 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.48e-01 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.gate_up_proj_weight_quantizers.161 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.gate_up_proj_weight_quantizers.162 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.15e-01 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.gate_up_proj_weight_quantizers.163 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.39e-01 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.gate_up_proj_weight_quantizers.164 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.gate_up_proj_weight_quantizers.165 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.73e-01 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.gate_up_proj_weight_quantizers.166 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.80e-01 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.gate_up_proj_weight_quantizers.167 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.gate_up_proj_weight_quantizers.168 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.gate_up_proj_weight_quantizers.169 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.gate_up_proj_weight_quantizers.170 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.03e-01 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.gate_up_proj_weight_quantizers.171 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.gate_up_proj_weight_quantizers.172 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.gate_up_proj_weight_quantizers.173 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.43e-01 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.gate_up_proj_weight_quantizers.174 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.gate_up_proj_weight_quantizers.175 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.39e-01 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.gate_up_proj_weight_quantizers.176 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.gate_up_proj_weight_quantizers.177 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.30e-01 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.gate_up_proj_weight_quantizers.178 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.gate_up_proj_weight_quantizers.179 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.gate_up_proj_weight_quantizers.180 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.51e-01 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.gate_up_proj_weight_quantizers.181 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.gate_up_proj_weight_quantizers.182 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.19e-01 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.gate_up_proj_weight_quantizers.183 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.86e-01 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.gate_up_proj_weight_quantizers.184 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.25e-01 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.gate_up_proj_weight_quantizers.185 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.gate_up_proj_weight_quantizers.186 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.gate_up_proj_weight_quantizers.187 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.gate_up_proj_weight_quantizers.188 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.65e-01 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.gate_up_proj_weight_quantizers.189 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.gate_up_proj_weight_quantizers.190 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.gate_up_proj_weight_quantizers.191 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.gate_up_proj_weight_quantizers.192 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.20e-01 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.gate_up_proj_weight_quantizers.193 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.gate_up_proj_weight_quantizers.194 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.48e-01 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.gate_up_proj_weight_quantizers.195 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.gate_up_proj_weight_quantizers.196 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.gate_up_proj_weight_quantizers.197 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.48e-01 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.gate_up_proj_weight_quantizers.198 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.13e-01 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.gate_up_proj_weight_quantizers.199 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.73e-01 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.gate_up_proj_weight_quantizers.200 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.14e-01 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.gate_up_proj_weight_quantizers.201 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.21e-01 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.gate_up_proj_weight_quantizers.202 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.66e-01 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.gate_up_proj_weight_quantizers.203 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.24e-01 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.gate_up_proj_weight_quantizers.204 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.gate_up_proj_weight_quantizers.205 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.15e-01 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.gate_up_proj_weight_quantizers.206 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.gate_up_proj_weight_quantizers.207 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.26e-01 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.gate_up_proj_weight_quantizers.208 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.gate_up_proj_weight_quantizers.209 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.44e-01 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.gate_up_proj_weight_quantizers.210 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.54e-01 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.gate_up_proj_weight_quantizers.211 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.gate_up_proj_weight_quantizers.212 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.29e-01 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.gate_up_proj_weight_quantizers.213 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.gate_up_proj_weight_quantizers.214 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.08e-01 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.gate_up_proj_weight_quantizers.215 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.23e-01 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.gate_up_proj_weight_quantizers.216 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.gate_up_proj_weight_quantizers.217 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.gate_up_proj_weight_quantizers.218 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.64e-01 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.gate_up_proj_weight_quantizers.219 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.gate_up_proj_weight_quantizers.220 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.gate_up_proj_weight_quantizers.221 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.08e-01 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.gate_up_proj_weight_quantizers.222 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.24e-01 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.gate_up_proj_weight_quantizers.223 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.gate_up_proj_weight_quantizers.224 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.18e-01 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.gate_up_proj_weight_quantizers.225 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.gate_up_proj_weight_quantizers.226 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.gate_up_proj_weight_quantizers.227 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.gate_up_proj_weight_quantizers.228 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.gate_up_proj_weight_quantizers.229 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.31e-01 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.gate_up_proj_weight_quantizers.230 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.26e-01 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.gate_up_proj_weight_quantizers.231 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.gate_up_proj_weight_quantizers.232 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.14e-01 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.gate_up_proj_weight_quantizers.233 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.gate_up_proj_weight_quantizers.234 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.36e-01 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.gate_up_proj_weight_quantizers.235 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.gate_up_proj_weight_quantizers.236 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.gate_up_proj_weight_quantizers.237 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.gate_up_proj_weight_quantizers.238 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.gate_up_proj_weight_quantizers.239 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.gate_up_proj_weight_quantizers.240 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.gate_up_proj_weight_quantizers.241 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.gate_up_proj_weight_quantizers.242 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.49e-01 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.gate_up_proj_weight_quantizers.243 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.33e-01 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.gate_up_proj_weight_quantizers.244 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.gate_up_proj_weight_quantizers.245 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.gate_up_proj_weight_quantizers.246 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.gate_up_proj_weight_quantizers.247 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.gate_up_proj_weight_quantizers.248 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.gate_up_proj_weight_quantizers.249 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.gate_up_proj_weight_quantizers.250 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.32e-01 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.gate_up_proj_weight_quantizers.251 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.24e-01 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.gate_up_proj_weight_quantizers.252 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.gate_up_proj_weight_quantizers.253 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.gate_up_proj_weight_quantizers.254 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.gate_up_proj_weight_quantizers.255 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.down_proj_input_quantizer TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.06e+01 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.down_proj_weight_quantizers.0 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.down_proj_weight_quantizers.1 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.down_proj_weight_quantizers.2 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.down_proj_weight_quantizers.3 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.down_proj_weight_quantizers.4 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.down_proj_weight_quantizers.5 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.22e-01 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.down_proj_weight_quantizers.6 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.64e-02 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.down_proj_weight_quantizers.7 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.61e-01 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.down_proj_weight_quantizers.8 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.33e-01 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.down_proj_weight_quantizers.9 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.down_proj_weight_quantizers.10 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.down_proj_weight_quantizers.11 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.53e-01 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.down_proj_weight_quantizers.12 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.12e-01 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.down_proj_weight_quantizers.13 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.down_proj_weight_quantizers.14 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.80e-01 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.down_proj_weight_quantizers.15 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.down_proj_weight_quantizers.16 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.down_proj_weight_quantizers.17 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.down_proj_weight_quantizers.18 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.down_proj_weight_quantizers.19 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.down_proj_weight_quantizers.20 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.down_proj_weight_quantizers.21 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.down_proj_weight_quantizers.22 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.down_proj_weight_quantizers.23 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.down_proj_weight_quantizers.24 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.down_proj_weight_quantizers.25 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.down_proj_weight_quantizers.26 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.64e-02 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.down_proj_weight_quantizers.27 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.06e-01 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.down_proj_weight_quantizers.28 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.down_proj_weight_quantizers.29 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.down_proj_weight_quantizers.30 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.down_proj_weight_quantizers.31 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.down_proj_weight_quantizers.32 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.28e-01 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.down_proj_weight_quantizers.33 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.down_proj_weight_quantizers.34 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.down_proj_weight_quantizers.35 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.50e-02 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.down_proj_weight_quantizers.36 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.down_proj_weight_quantizers.37 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.60e-01 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.down_proj_weight_quantizers.38 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.35e-01 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.down_proj_weight_quantizers.39 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.92e-01 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.down_proj_weight_quantizers.40 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.91e-01 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.down_proj_weight_quantizers.41 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.29e-01 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.down_proj_weight_quantizers.42 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.down_proj_weight_quantizers.43 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.down_proj_weight_quantizers.44 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.down_proj_weight_quantizers.45 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.down_proj_weight_quantizers.46 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.22e-01 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.down_proj_weight_quantizers.47 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.down_proj_weight_quantizers.48 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.down_proj_weight_quantizers.49 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.down_proj_weight_quantizers.50 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.58e-01 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.down_proj_weight_quantizers.51 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.64e-02 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.down_proj_weight_quantizers.52 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.down_proj_weight_quantizers.53 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.92e-01 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.down_proj_weight_quantizers.54 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.30e-01 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.down_proj_weight_quantizers.55 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.54e-02 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.down_proj_weight_quantizers.56 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.down_proj_weight_quantizers.57 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.down_proj_weight_quantizers.58 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.down_proj_weight_quantizers.59 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.down_proj_weight_quantizers.60 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.down_proj_weight_quantizers.61 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.down_proj_weight_quantizers.62 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.32e-01 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.down_proj_weight_quantizers.63 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.down_proj_weight_quantizers.64 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.33e-01 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.down_proj_weight_quantizers.65 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.down_proj_weight_quantizers.66 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.down_proj_weight_quantizers.67 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.down_proj_weight_quantizers.68 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.down_proj_weight_quantizers.69 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.15e-02 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.down_proj_weight_quantizers.70 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.down_proj_weight_quantizers.71 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.29e-01 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.down_proj_weight_quantizers.72 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.down_proj_weight_quantizers.73 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.down_proj_weight_quantizers.74 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.45e-02 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.down_proj_weight_quantizers.75 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.down_proj_weight_quantizers.76 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.68e-01 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.down_proj_weight_quantizers.77 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.down_proj_weight_quantizers.78 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.down_proj_weight_quantizers.79 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.down_proj_weight_quantizers.80 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.50e-02 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.down_proj_weight_quantizers.81 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.down_proj_weight_quantizers.82 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.down_proj_weight_quantizers.83 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.down_proj_weight_quantizers.84 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.down_proj_weight_quantizers.85 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.down_proj_weight_quantizers.86 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.down_proj_weight_quantizers.87 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.down_proj_weight_quantizers.88 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.down_proj_weight_quantizers.89 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.down_proj_weight_quantizers.90 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.down_proj_weight_quantizers.91 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.down_proj_weight_quantizers.92 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.67e-01 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.down_proj_weight_quantizers.93 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.down_proj_weight_quantizers.94 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.down_proj_weight_quantizers.95 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.58e-01 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.down_proj_weight_quantizers.96 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.down_proj_weight_quantizers.97 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.down_proj_weight_quantizers.98 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.down_proj_weight_quantizers.99 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.down_proj_weight_quantizers.100 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.50e-02 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.down_proj_weight_quantizers.101 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.down_proj_weight_quantizers.102 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.77e-01 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.down_proj_weight_quantizers.103 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.down_proj_weight_quantizers.104 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.down_proj_weight_quantizers.105 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.down_proj_weight_quantizers.106 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.down_proj_weight_quantizers.107 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.down_proj_weight_quantizers.108 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.58e-01 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.down_proj_weight_quantizers.109 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.down_proj_weight_quantizers.110 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.28e-01 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.down_proj_weight_quantizers.111 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.down_proj_weight_quantizers.112 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.13e-01 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.down_proj_weight_quantizers.113 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.down_proj_weight_quantizers.114 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.43e-01 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.down_proj_weight_quantizers.115 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.down_proj_weight_quantizers.116 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.78e-01 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.down_proj_weight_quantizers.117 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.down_proj_weight_quantizers.118 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.down_proj_weight_quantizers.119 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.50e-02 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.down_proj_weight_quantizers.120 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.down_proj_weight_quantizers.121 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.down_proj_weight_quantizers.122 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.59e-02 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.down_proj_weight_quantizers.123 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.79e-01 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.down_proj_weight_quantizers.124 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.54e-02 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.down_proj_weight_quantizers.125 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.down_proj_weight_quantizers.126 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.down_proj_weight_quantizers.127 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.down_proj_weight_quantizers.128 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.14e-01 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.down_proj_weight_quantizers.129 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.down_proj_weight_quantizers.130 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.down_proj_weight_quantizers.131 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.45e-02 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.down_proj_weight_quantizers.132 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.down_proj_weight_quantizers.133 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.54e-01 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.down_proj_weight_quantizers.134 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.down_proj_weight_quantizers.135 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.down_proj_weight_quantizers.136 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.39e-01 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.down_proj_weight_quantizers.137 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.down_proj_weight_quantizers.138 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.59e-02 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.down_proj_weight_quantizers.139 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.down_proj_weight_quantizers.140 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.down_proj_weight_quantizers.141 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.down_proj_weight_quantizers.142 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.down_proj_weight_quantizers.143 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.43e-01 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.down_proj_weight_quantizers.144 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.35e-02 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.down_proj_weight_quantizers.145 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.59e-02 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.down_proj_weight_quantizers.146 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.81e-01 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.down_proj_weight_quantizers.147 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.down_proj_weight_quantizers.148 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.14e-01 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.down_proj_weight_quantizers.149 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.down_proj_weight_quantizers.150 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.down_proj_weight_quantizers.151 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.90e-01 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.down_proj_weight_quantizers.152 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.down_proj_weight_quantizers.153 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.down_proj_weight_quantizers.154 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.down_proj_weight_quantizers.155 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.down_proj_weight_quantizers.156 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.down_proj_weight_quantizers.157 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.down_proj_weight_quantizers.158 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.16e-01 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.down_proj_weight_quantizers.159 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.23e-01 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.down_proj_weight_quantizers.160 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.down_proj_weight_quantizers.161 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.down_proj_weight_quantizers.162 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.down_proj_weight_quantizers.163 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.31e-01 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.down_proj_weight_quantizers.164 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.down_proj_weight_quantizers.165 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.down_proj_weight_quantizers.166 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.down_proj_weight_quantizers.167 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.down_proj_weight_quantizers.168 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.down_proj_weight_quantizers.169 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.down_proj_weight_quantizers.170 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.64e-01 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.down_proj_weight_quantizers.171 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.down_proj_weight_quantizers.172 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.down_proj_weight_quantizers.173 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.30e-01 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.down_proj_weight_quantizers.174 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.down_proj_weight_quantizers.175 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.down_proj_weight_quantizers.176 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.down_proj_weight_quantizers.177 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.down_proj_weight_quantizers.178 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.down_proj_weight_quantizers.179 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.down_proj_weight_quantizers.180 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.95e-01 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.down_proj_weight_quantizers.181 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.down_proj_weight_quantizers.182 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.79e-01 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.down_proj_weight_quantizers.183 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.down_proj_weight_quantizers.184 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.down_proj_weight_quantizers.185 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.down_proj_weight_quantizers.186 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.down_proj_weight_quantizers.187 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.down_proj_weight_quantizers.188 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.78e-01 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.down_proj_weight_quantizers.189 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.down_proj_weight_quantizers.190 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.down_proj_weight_quantizers.191 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.down_proj_weight_quantizers.192 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.42e-01 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.down_proj_weight_quantizers.193 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.down_proj_weight_quantizers.194 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.43e-01 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.down_proj_weight_quantizers.195 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.down_proj_weight_quantizers.196 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.down_proj_weight_quantizers.197 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=6.25e-01 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.down_proj_weight_quantizers.198 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.down_proj_weight_quantizers.199 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.19e-01 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.down_proj_weight_quantizers.200 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.down_proj_weight_quantizers.201 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=4.57e-01 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.down_proj_weight_quantizers.202 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=4.80e-01 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.down_proj_weight_quantizers.203 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.down_proj_weight_quantizers.204 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.down_proj_weight_quantizers.205 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.down_proj_weight_quantizers.206 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.26e-01 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.down_proj_weight_quantizers.207 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.down_proj_weight_quantizers.208 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.down_proj_weight_quantizers.209 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.90e-01 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.down_proj_weight_quantizers.210 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.91e-01 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.down_proj_weight_quantizers.211 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.down_proj_weight_quantizers.212 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.45e-02 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.down_proj_weight_quantizers.213 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.down_proj_weight_quantizers.214 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.50e-01 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.down_proj_weight_quantizers.215 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.down_proj_weight_quantizers.216 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.99e-01 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.down_proj_weight_quantizers.217 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.64e-02 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.down_proj_weight_quantizers.218 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.down_proj_weight_quantizers.219 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.down_proj_weight_quantizers.220 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.down_proj_weight_quantizers.221 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.78e-01 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.down_proj_weight_quantizers.222 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.26e-01 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.down_proj_weight_quantizers.223 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.down_proj_weight_quantizers.224 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.down_proj_weight_quantizers.225 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.down_proj_weight_quantizers.226 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.down_proj_weight_quantizers.227 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.down_proj_weight_quantizers.228 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.54e-02 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.down_proj_weight_quantizers.229 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.64e-02 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.down_proj_weight_quantizers.230 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.20e-01 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.down_proj_weight_quantizers.231 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.down_proj_weight_quantizers.232 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.down_proj_weight_quantizers.233 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.down_proj_weight_quantizers.234 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.43e-01 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.down_proj_weight_quantizers.235 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.down_proj_weight_quantizers.236 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.down_proj_weight_quantizers.237 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.down_proj_weight_quantizers.238 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.down_proj_weight_quantizers.239 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.35e-02 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.down_proj_weight_quantizers.240 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.down_proj_weight_quantizers.241 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.down_proj_weight_quantizers.242 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.down_proj_weight_quantizers.243 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.89e-01 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.down_proj_weight_quantizers.244 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.down_proj_weight_quantizers.245 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.down_proj_weight_quantizers.246 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.down_proj_weight_quantizers.247 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.down_proj_weight_quantizers.248 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.54e-02 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.down_proj_weight_quantizers.249 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.down_proj_weight_quantizers.250 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.down_proj_weight_quantizers.251 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.64e-02 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.down_proj_weight_quantizers.252 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.down_proj_weight_quantizers.253 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.down_proj_weight_quantizers.254 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.down_proj_weight_quantizers.255 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.49.mlp.shared_experts.gate_proj.input_quantizer TensorQuantizer(disabled) +model.layers.49.mlp.shared_experts.gate_proj.output_quantizer TensorQuantizer(disabled) +model.layers.49.mlp.shared_experts.gate_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.49.mlp.shared_experts.up_proj.input_quantizer TensorQuantizer(disabled) +model.layers.49.mlp.shared_experts.up_proj.output_quantizer TensorQuantizer(disabled) +model.layers.49.mlp.shared_experts.up_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.49.mlp.shared_experts.down_proj.input_quantizer TensorQuantizer(disabled) +model.layers.49.mlp.shared_experts.down_proj.output_quantizer TensorQuantizer(disabled) +model.layers.49.mlp.shared_experts.down_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.50.self_attn.q_a_proj.input_quantizer TensorQuantizer(disabled) +model.layers.50.self_attn.q_a_proj.output_quantizer TensorQuantizer(disabled) +model.layers.50.self_attn.q_a_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.50.self_attn.q_b_proj.input_quantizer TensorQuantizer(disabled) +model.layers.50.self_attn.q_b_proj.output_quantizer TensorQuantizer(disabled) +model.layers.50.self_attn.q_b_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.50.self_attn.kv_a_proj_with_mqa.input_quantizer TensorQuantizer(disabled) +model.layers.50.self_attn.kv_a_proj_with_mqa.output_quantizer TensorQuantizer(disabled) +model.layers.50.self_attn.kv_a_proj_with_mqa.weight_quantizer TensorQuantizer(disabled) +model.layers.50.self_attn.kv_b_proj.input_quantizer TensorQuantizer(disabled) +model.layers.50.self_attn.kv_b_proj.output_quantizer TensorQuantizer(disabled) +model.layers.50.self_attn.kv_b_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.50.self_attn.o_proj.input_quantizer TensorQuantizer(disabled) +model.layers.50.self_attn.o_proj.output_quantizer TensorQuantizer(disabled) +model.layers.50.self_attn.o_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.50.self_attn.indexer.wq_b.input_quantizer TensorQuantizer(disabled) +model.layers.50.self_attn.indexer.wq_b.output_quantizer TensorQuantizer(disabled) +model.layers.50.self_attn.indexer.wq_b.weight_quantizer TensorQuantizer(disabled) +model.layers.50.self_attn.indexer.wk.input_quantizer TensorQuantizer(disabled) +model.layers.50.self_attn.indexer.wk.output_quantizer TensorQuantizer(disabled) +model.layers.50.self_attn.indexer.wk.weight_quantizer TensorQuantizer(disabled) +model.layers.50.self_attn.indexer.k_norm.input_quantizer TensorQuantizer(disabled) +model.layers.50.self_attn.indexer.k_norm.output_quantizer TensorQuantizer(disabled) +model.layers.50.self_attn.indexer.weights_proj.input_quantizer TensorQuantizer(disabled) +model.layers.50.self_attn.indexer.weights_proj.output_quantizer TensorQuantizer(disabled) +model.layers.50.self_attn.indexer.weights_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.50.self_attn.q_bmm_quantizer TensorQuantizer(disabled) +model.layers.50.self_attn.k_bmm_quantizer TensorQuantizer((4, 3) bit fake per-tensor amax=4.48e+02(const) calibrator=MaxCalibrator quant) +model.layers.50.self_attn.v_bmm_quantizer TensorQuantizer((4, 3) bit fake per-tensor amax=4.48e+02(const) calibrator=MaxCalibrator quant) +model.layers.50.self_attn.p_bmm_quantizer TensorQuantizer(disabled) +model.layers.50.mlp.experts.gate_up_proj_input_quantizer TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.75e+00 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.gate_up_proj_weight_quantizers.0 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.65e-01 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.gate_up_proj_weight_quantizers.1 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.13e-01 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.gate_up_proj_weight_quantizers.2 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.gate_up_proj_weight_quantizers.3 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.gate_up_proj_weight_quantizers.4 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.gate_up_proj_weight_quantizers.5 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.gate_up_proj_weight_quantizers.6 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.gate_up_proj_weight_quantizers.7 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.gate_up_proj_weight_quantizers.8 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.gate_up_proj_weight_quantizers.9 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.gate_up_proj_weight_quantizers.10 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.68e-01 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.gate_up_proj_weight_quantizers.11 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.gate_up_proj_weight_quantizers.12 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.46e-01 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.gate_up_proj_weight_quantizers.13 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.29e-01 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.gate_up_proj_weight_quantizers.14 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.34e-01 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.gate_up_proj_weight_quantizers.15 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.gate_up_proj_weight_quantizers.16 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.gate_up_proj_weight_quantizers.17 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.25e-01 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.gate_up_proj_weight_quantizers.18 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.gate_up_proj_weight_quantizers.19 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.17e-01 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.gate_up_proj_weight_quantizers.20 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.gate_up_proj_weight_quantizers.21 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.gate_up_proj_weight_quantizers.22 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.15e-01 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.gate_up_proj_weight_quantizers.23 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.gate_up_proj_weight_quantizers.24 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.gate_up_proj_weight_quantizers.25 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.gate_up_proj_weight_quantizers.26 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.22e-01 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.gate_up_proj_weight_quantizers.27 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.15e-01 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.gate_up_proj_weight_quantizers.28 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.17e-01 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.gate_up_proj_weight_quantizers.29 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.gate_up_proj_weight_quantizers.30 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.gate_up_proj_weight_quantizers.31 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.gate_up_proj_weight_quantizers.32 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.gate_up_proj_weight_quantizers.33 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.gate_up_proj_weight_quantizers.34 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.gate_up_proj_weight_quantizers.35 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.gate_up_proj_weight_quantizers.36 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.gate_up_proj_weight_quantizers.37 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.60e-01 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.gate_up_proj_weight_quantizers.38 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.gate_up_proj_weight_quantizers.39 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.15e-01 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.gate_up_proj_weight_quantizers.40 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.28e-01 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.gate_up_proj_weight_quantizers.41 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.gate_up_proj_weight_quantizers.42 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.gate_up_proj_weight_quantizers.43 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.35e-01 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.gate_up_proj_weight_quantizers.44 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.gate_up_proj_weight_quantizers.45 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.gate_up_proj_weight_quantizers.46 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.16e-01 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.gate_up_proj_weight_quantizers.47 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.gate_up_proj_weight_quantizers.48 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.gate_up_proj_weight_quantizers.49 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.53e-01 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.gate_up_proj_weight_quantizers.50 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.gate_up_proj_weight_quantizers.51 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.34e-01 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.gate_up_proj_weight_quantizers.52 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.gate_up_proj_weight_quantizers.53 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.gate_up_proj_weight_quantizers.54 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.gate_up_proj_weight_quantizers.55 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.gate_up_proj_weight_quantizers.56 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.gate_up_proj_weight_quantizers.57 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.gate_up_proj_weight_quantizers.58 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.gate_up_proj_weight_quantizers.59 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.gate_up_proj_weight_quantizers.60 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.gate_up_proj_weight_quantizers.61 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.gate_up_proj_weight_quantizers.62 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.66e-01 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.gate_up_proj_weight_quantizers.63 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.99e-01 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.gate_up_proj_weight_quantizers.64 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.gate_up_proj_weight_quantizers.65 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.13e-01 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.gate_up_proj_weight_quantizers.66 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.14e-01 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.gate_up_proj_weight_quantizers.67 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.69e-01 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.gate_up_proj_weight_quantizers.68 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.gate_up_proj_weight_quantizers.69 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.gate_up_proj_weight_quantizers.70 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.gate_up_proj_weight_quantizers.71 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.gate_up_proj_weight_quantizers.72 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.gate_up_proj_weight_quantizers.73 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.gate_up_proj_weight_quantizers.74 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.gate_up_proj_weight_quantizers.75 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.gate_up_proj_weight_quantizers.76 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.gate_up_proj_weight_quantizers.77 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.gate_up_proj_weight_quantizers.78 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.20e-01 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.gate_up_proj_weight_quantizers.79 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.19e-01 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.gate_up_proj_weight_quantizers.80 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.gate_up_proj_weight_quantizers.81 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.gate_up_proj_weight_quantizers.82 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.gate_up_proj_weight_quantizers.83 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.gate_up_proj_weight_quantizers.84 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.gate_up_proj_weight_quantizers.85 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.28e-01 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.gate_up_proj_weight_quantizers.86 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.gate_up_proj_weight_quantizers.87 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.69e-01 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.gate_up_proj_weight_quantizers.88 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.98e-01 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.gate_up_proj_weight_quantizers.89 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.gate_up_proj_weight_quantizers.90 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.gate_up_proj_weight_quantizers.91 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.gate_up_proj_weight_quantizers.92 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.53e-01 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.gate_up_proj_weight_quantizers.93 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.gate_up_proj_weight_quantizers.94 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.gate_up_proj_weight_quantizers.95 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.25e-01 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.gate_up_proj_weight_quantizers.96 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.gate_up_proj_weight_quantizers.97 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.gate_up_proj_weight_quantizers.98 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.20e-01 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.gate_up_proj_weight_quantizers.99 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.gate_up_proj_weight_quantizers.100 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.15e-01 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.gate_up_proj_weight_quantizers.101 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.54e-01 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.gate_up_proj_weight_quantizers.102 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.gate_up_proj_weight_quantizers.103 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.gate_up_proj_weight_quantizers.104 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.30e-01 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.gate_up_proj_weight_quantizers.105 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.gate_up_proj_weight_quantizers.106 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.gate_up_proj_weight_quantizers.107 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.21e-01 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.gate_up_proj_weight_quantizers.108 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.gate_up_proj_weight_quantizers.109 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=4.06e-01 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.gate_up_proj_weight_quantizers.110 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.gate_up_proj_weight_quantizers.111 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.gate_up_proj_weight_quantizers.112 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.gate_up_proj_weight_quantizers.113 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.14e-01 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.gate_up_proj_weight_quantizers.114 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.39e-01 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.gate_up_proj_weight_quantizers.115 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.gate_up_proj_weight_quantizers.116 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.gate_up_proj_weight_quantizers.117 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.gate_up_proj_weight_quantizers.118 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.gate_up_proj_weight_quantizers.119 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.22e-01 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.gate_up_proj_weight_quantizers.120 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.gate_up_proj_weight_quantizers.121 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.gate_up_proj_weight_quantizers.122 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.25e-01 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.gate_up_proj_weight_quantizers.123 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.gate_up_proj_weight_quantizers.124 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.gate_up_proj_weight_quantizers.125 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.23e-01 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.gate_up_proj_weight_quantizers.126 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.gate_up_proj_weight_quantizers.127 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.gate_up_proj_weight_quantizers.128 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.gate_up_proj_weight_quantizers.129 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.gate_up_proj_weight_quantizers.130 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.gate_up_proj_weight_quantizers.131 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.21e-01 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.gate_up_proj_weight_quantizers.132 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.gate_up_proj_weight_quantizers.133 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.gate_up_proj_weight_quantizers.134 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.gate_up_proj_weight_quantizers.135 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.gate_up_proj_weight_quantizers.136 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.gate_up_proj_weight_quantizers.137 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.gate_up_proj_weight_quantizers.138 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.gate_up_proj_weight_quantizers.139 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.gate_up_proj_weight_quantizers.140 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.gate_up_proj_weight_quantizers.141 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.22e-01 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.gate_up_proj_weight_quantizers.142 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.53e-01 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.gate_up_proj_weight_quantizers.143 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.gate_up_proj_weight_quantizers.144 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.44e-01 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.gate_up_proj_weight_quantizers.145 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.gate_up_proj_weight_quantizers.146 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.42e-01 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.gate_up_proj_weight_quantizers.147 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.14e-01 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.gate_up_proj_weight_quantizers.148 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.gate_up_proj_weight_quantizers.149 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.gate_up_proj_weight_quantizers.150 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.gate_up_proj_weight_quantizers.151 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.17e-01 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.gate_up_proj_weight_quantizers.152 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.gate_up_proj_weight_quantizers.153 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.gate_up_proj_weight_quantizers.154 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.gate_up_proj_weight_quantizers.155 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.gate_up_proj_weight_quantizers.156 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.gate_up_proj_weight_quantizers.157 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.14e-01 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.gate_up_proj_weight_quantizers.158 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.gate_up_proj_weight_quantizers.159 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.gate_up_proj_weight_quantizers.160 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.gate_up_proj_weight_quantizers.161 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.gate_up_proj_weight_quantizers.162 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.gate_up_proj_weight_quantizers.163 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.15e-01 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.gate_up_proj_weight_quantizers.164 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.gate_up_proj_weight_quantizers.165 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.29e-01 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.gate_up_proj_weight_quantizers.166 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.gate_up_proj_weight_quantizers.167 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.gate_up_proj_weight_quantizers.168 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.gate_up_proj_weight_quantizers.169 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.gate_up_proj_weight_quantizers.170 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.gate_up_proj_weight_quantizers.171 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.20e-01 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.gate_up_proj_weight_quantizers.172 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.gate_up_proj_weight_quantizers.173 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.gate_up_proj_weight_quantizers.174 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.gate_up_proj_weight_quantizers.175 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.15e-01 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.gate_up_proj_weight_quantizers.176 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.gate_up_proj_weight_quantizers.177 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.34e-01 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.gate_up_proj_weight_quantizers.178 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.gate_up_proj_weight_quantizers.179 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.gate_up_proj_weight_quantizers.180 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.gate_up_proj_weight_quantizers.181 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.73e-01 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.gate_up_proj_weight_quantizers.182 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.gate_up_proj_weight_quantizers.183 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.27e-01 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.gate_up_proj_weight_quantizers.184 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.gate_up_proj_weight_quantizers.185 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.gate_up_proj_weight_quantizers.186 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.gate_up_proj_weight_quantizers.187 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.16e-01 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.gate_up_proj_weight_quantizers.188 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.gate_up_proj_weight_quantizers.189 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.gate_up_proj_weight_quantizers.190 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.04e-01 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.gate_up_proj_weight_quantizers.191 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.60e-01 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.gate_up_proj_weight_quantizers.192 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.gate_up_proj_weight_quantizers.193 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.31e-01 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.gate_up_proj_weight_quantizers.194 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.gate_up_proj_weight_quantizers.195 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.gate_up_proj_weight_quantizers.196 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.gate_up_proj_weight_quantizers.197 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.gate_up_proj_weight_quantizers.198 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.gate_up_proj_weight_quantizers.199 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.27e-01 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.gate_up_proj_weight_quantizers.200 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.gate_up_proj_weight_quantizers.201 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.gate_up_proj_weight_quantizers.202 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.gate_up_proj_weight_quantizers.203 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.gate_up_proj_weight_quantizers.204 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.gate_up_proj_weight_quantizers.205 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.gate_up_proj_weight_quantizers.206 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.gate_up_proj_weight_quantizers.207 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.gate_up_proj_weight_quantizers.208 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.gate_up_proj_weight_quantizers.209 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.gate_up_proj_weight_quantizers.210 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.23e-01 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.gate_up_proj_weight_quantizers.211 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.89e-01 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.gate_up_proj_weight_quantizers.212 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.gate_up_proj_weight_quantizers.213 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.gate_up_proj_weight_quantizers.214 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.19e-01 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.gate_up_proj_weight_quantizers.215 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.gate_up_proj_weight_quantizers.216 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.gate_up_proj_weight_quantizers.217 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.25e-01 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.gate_up_proj_weight_quantizers.218 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.gate_up_proj_weight_quantizers.219 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.gate_up_proj_weight_quantizers.220 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.gate_up_proj_weight_quantizers.221 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.gate_up_proj_weight_quantizers.222 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.gate_up_proj_weight_quantizers.223 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.gate_up_proj_weight_quantizers.224 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.gate_up_proj_weight_quantizers.225 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.gate_up_proj_weight_quantizers.226 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.gate_up_proj_weight_quantizers.227 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.gate_up_proj_weight_quantizers.228 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.gate_up_proj_weight_quantizers.229 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.17e-01 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.gate_up_proj_weight_quantizers.230 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.13e-01 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.gate_up_proj_weight_quantizers.231 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.gate_up_proj_weight_quantizers.232 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.gate_up_proj_weight_quantizers.233 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.gate_up_proj_weight_quantizers.234 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.gate_up_proj_weight_quantizers.235 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.83e-01 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.gate_up_proj_weight_quantizers.236 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.gate_up_proj_weight_quantizers.237 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.gate_up_proj_weight_quantizers.238 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.gate_up_proj_weight_quantizers.239 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.60e-01 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.gate_up_proj_weight_quantizers.240 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.gate_up_proj_weight_quantizers.241 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.gate_up_proj_weight_quantizers.242 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.gate_up_proj_weight_quantizers.243 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.gate_up_proj_weight_quantizers.244 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.25e-01 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.gate_up_proj_weight_quantizers.245 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.21e-01 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.gate_up_proj_weight_quantizers.246 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.gate_up_proj_weight_quantizers.247 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.gate_up_proj_weight_quantizers.248 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.15e-01 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.gate_up_proj_weight_quantizers.249 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.25e-01 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.gate_up_proj_weight_quantizers.250 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.19e-01 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.gate_up_proj_weight_quantizers.251 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.gate_up_proj_weight_quantizers.252 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.gate_up_proj_weight_quantizers.253 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.gate_up_proj_weight_quantizers.254 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.gate_up_proj_weight_quantizers.255 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.down_proj_input_quantizer TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=4.52e+01 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.down_proj_weight_quantizers.0 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.19e-01 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.down_proj_weight_quantizers.1 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.down_proj_weight_quantizers.2 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.down_proj_weight_quantizers.3 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.down_proj_weight_quantizers.4 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.down_proj_weight_quantizers.5 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.80e-01 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.down_proj_weight_quantizers.6 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.down_proj_weight_quantizers.7 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.down_proj_weight_quantizers.8 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.down_proj_weight_quantizers.9 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.down_proj_weight_quantizers.10 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.22e-01 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.down_proj_weight_quantizers.11 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.down_proj_weight_quantizers.12 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.down_proj_weight_quantizers.13 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.41e-01 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.down_proj_weight_quantizers.14 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.50e-02 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.down_proj_weight_quantizers.15 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.down_proj_weight_quantizers.16 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.down_proj_weight_quantizers.17 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.down_proj_weight_quantizers.18 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.50e-02 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.down_proj_weight_quantizers.19 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.43e-01 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.down_proj_weight_quantizers.20 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.down_proj_weight_quantizers.21 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.down_proj_weight_quantizers.22 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.down_proj_weight_quantizers.23 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.down_proj_weight_quantizers.24 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.45e-02 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.down_proj_weight_quantizers.25 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.down_proj_weight_quantizers.26 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.down_proj_weight_quantizers.27 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.down_proj_weight_quantizers.28 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.down_proj_weight_quantizers.29 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.down_proj_weight_quantizers.30 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.down_proj_weight_quantizers.31 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.down_proj_weight_quantizers.32 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.down_proj_weight_quantizers.33 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.down_proj_weight_quantizers.34 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.down_proj_weight_quantizers.35 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.down_proj_weight_quantizers.36 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.down_proj_weight_quantizers.37 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.down_proj_weight_quantizers.38 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.down_proj_weight_quantizers.39 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.down_proj_weight_quantizers.40 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.down_proj_weight_quantizers.41 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.down_proj_weight_quantizers.42 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.down_proj_weight_quantizers.43 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.50e-02 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.down_proj_weight_quantizers.44 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.down_proj_weight_quantizers.45 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.down_proj_weight_quantizers.46 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.41e-01 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.down_proj_weight_quantizers.47 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.down_proj_weight_quantizers.48 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.down_proj_weight_quantizers.49 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.down_proj_weight_quantizers.50 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.29e-01 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.down_proj_weight_quantizers.51 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.down_proj_weight_quantizers.52 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.down_proj_weight_quantizers.53 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.down_proj_weight_quantizers.54 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.50e-02 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.down_proj_weight_quantizers.55 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.down_proj_weight_quantizers.56 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.down_proj_weight_quantizers.57 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.down_proj_weight_quantizers.58 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.down_proj_weight_quantizers.59 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.82e-01 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.down_proj_weight_quantizers.60 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.down_proj_weight_quantizers.61 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.down_proj_weight_quantizers.62 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.down_proj_weight_quantizers.63 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.48e-01 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.down_proj_weight_quantizers.64 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.54e-02 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.down_proj_weight_quantizers.65 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.down_proj_weight_quantizers.66 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.down_proj_weight_quantizers.67 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.down_proj_weight_quantizers.68 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.50e-02 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.down_proj_weight_quantizers.69 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.down_proj_weight_quantizers.70 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.down_proj_weight_quantizers.71 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.down_proj_weight_quantizers.72 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.down_proj_weight_quantizers.73 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.down_proj_weight_quantizers.74 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.down_proj_weight_quantizers.75 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.down_proj_weight_quantizers.76 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.down_proj_weight_quantizers.77 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.down_proj_weight_quantizers.78 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.down_proj_weight_quantizers.79 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.down_proj_weight_quantizers.80 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.down_proj_weight_quantizers.81 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.down_proj_weight_quantizers.82 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.down_proj_weight_quantizers.83 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.13e-01 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.down_proj_weight_quantizers.84 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.down_proj_weight_quantizers.85 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.down_proj_weight_quantizers.86 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.86e-01 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.down_proj_weight_quantizers.87 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.47e-01 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.down_proj_weight_quantizers.88 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.67e-01 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.down_proj_weight_quantizers.89 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.down_proj_weight_quantizers.90 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.down_proj_weight_quantizers.91 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.down_proj_weight_quantizers.92 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.34e-01 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.down_proj_weight_quantizers.93 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.64e-02 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.down_proj_weight_quantizers.94 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.down_proj_weight_quantizers.95 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.24e-01 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.down_proj_weight_quantizers.96 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.down_proj_weight_quantizers.97 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.down_proj_weight_quantizers.98 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.25e-01 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.down_proj_weight_quantizers.99 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.down_proj_weight_quantizers.100 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.down_proj_weight_quantizers.101 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.down_proj_weight_quantizers.102 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.down_proj_weight_quantizers.103 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.down_proj_weight_quantizers.104 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.down_proj_weight_quantizers.105 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.down_proj_weight_quantizers.106 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.down_proj_weight_quantizers.107 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.down_proj_weight_quantizers.108 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.down_proj_weight_quantizers.109 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.40e-01 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.down_proj_weight_quantizers.110 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.down_proj_weight_quantizers.111 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.down_proj_weight_quantizers.112 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.down_proj_weight_quantizers.113 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.down_proj_weight_quantizers.114 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.down_proj_weight_quantizers.115 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.down_proj_weight_quantizers.116 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.54e-02 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.down_proj_weight_quantizers.117 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.down_proj_weight_quantizers.118 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.down_proj_weight_quantizers.119 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.down_proj_weight_quantizers.120 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.down_proj_weight_quantizers.121 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.down_proj_weight_quantizers.122 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.26e-01 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.down_proj_weight_quantizers.123 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.down_proj_weight_quantizers.124 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.down_proj_weight_quantizers.125 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.down_proj_weight_quantizers.126 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.down_proj_weight_quantizers.127 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.down_proj_weight_quantizers.128 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.down_proj_weight_quantizers.129 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.down_proj_weight_quantizers.130 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.down_proj_weight_quantizers.131 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.41e-01 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.down_proj_weight_quantizers.132 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.down_proj_weight_quantizers.133 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.down_proj_weight_quantizers.134 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.down_proj_weight_quantizers.135 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.down_proj_weight_quantizers.136 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.down_proj_weight_quantizers.137 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.35e-02 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.down_proj_weight_quantizers.138 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.down_proj_weight_quantizers.139 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.down_proj_weight_quantizers.140 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.down_proj_weight_quantizers.141 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.30e-02 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.down_proj_weight_quantizers.142 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.down_proj_weight_quantizers.143 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.down_proj_weight_quantizers.144 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.64e-02 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.down_proj_weight_quantizers.145 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.down_proj_weight_quantizers.146 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.14e-01 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.down_proj_weight_quantizers.147 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.down_proj_weight_quantizers.148 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.down_proj_weight_quantizers.149 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.down_proj_weight_quantizers.150 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.39e-01 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.down_proj_weight_quantizers.151 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.43e-01 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.down_proj_weight_quantizers.152 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.down_proj_weight_quantizers.153 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.down_proj_weight_quantizers.154 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.down_proj_weight_quantizers.155 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.down_proj_weight_quantizers.156 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.down_proj_weight_quantizers.157 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.23e-01 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.down_proj_weight_quantizers.158 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.down_proj_weight_quantizers.159 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.down_proj_weight_quantizers.160 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.down_proj_weight_quantizers.161 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.down_proj_weight_quantizers.162 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.down_proj_weight_quantizers.163 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.down_proj_weight_quantizers.164 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.down_proj_weight_quantizers.165 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.down_proj_weight_quantizers.166 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.down_proj_weight_quantizers.167 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.down_proj_weight_quantizers.168 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.down_proj_weight_quantizers.169 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.35e-02 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.down_proj_weight_quantizers.170 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.down_proj_weight_quantizers.171 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.down_proj_weight_quantizers.172 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.down_proj_weight_quantizers.173 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.97e-01 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.down_proj_weight_quantizers.174 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.down_proj_weight_quantizers.175 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.17e-01 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.down_proj_weight_quantizers.176 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.down_proj_weight_quantizers.177 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.down_proj_weight_quantizers.178 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.down_proj_weight_quantizers.179 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.down_proj_weight_quantizers.180 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.down_proj_weight_quantizers.181 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.down_proj_weight_quantizers.182 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.down_proj_weight_quantizers.183 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.56e-01 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.down_proj_weight_quantizers.184 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.down_proj_weight_quantizers.185 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.down_proj_weight_quantizers.186 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.down_proj_weight_quantizers.187 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.down_proj_weight_quantizers.188 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.down_proj_weight_quantizers.189 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.down_proj_weight_quantizers.190 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.down_proj_weight_quantizers.191 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.down_proj_weight_quantizers.192 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.down_proj_weight_quantizers.193 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.54e-02 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.down_proj_weight_quantizers.194 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.down_proj_weight_quantizers.195 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.down_proj_weight_quantizers.196 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.down_proj_weight_quantizers.197 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.64e-02 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.down_proj_weight_quantizers.198 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.down_proj_weight_quantizers.199 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.down_proj_weight_quantizers.200 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.down_proj_weight_quantizers.201 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.down_proj_weight_quantizers.202 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.down_proj_weight_quantizers.203 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.down_proj_weight_quantizers.204 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.down_proj_weight_quantizers.205 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.down_proj_weight_quantizers.206 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.down_proj_weight_quantizers.207 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.down_proj_weight_quantizers.208 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.down_proj_weight_quantizers.209 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.down_proj_weight_quantizers.210 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.down_proj_weight_quantizers.211 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.31e-01 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.down_proj_weight_quantizers.212 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.down_proj_weight_quantizers.213 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.down_proj_weight_quantizers.214 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.70e-01 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.down_proj_weight_quantizers.215 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.down_proj_weight_quantizers.216 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.30e-02 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.down_proj_weight_quantizers.217 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.down_proj_weight_quantizers.218 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.down_proj_weight_quantizers.219 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.64e-02 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.down_proj_weight_quantizers.220 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.down_proj_weight_quantizers.221 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.down_proj_weight_quantizers.222 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.down_proj_weight_quantizers.223 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.down_proj_weight_quantizers.224 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.down_proj_weight_quantizers.225 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.down_proj_weight_quantizers.226 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.down_proj_weight_quantizers.227 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.down_proj_weight_quantizers.228 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.down_proj_weight_quantizers.229 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.down_proj_weight_quantizers.230 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.down_proj_weight_quantizers.231 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.62e-01 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.down_proj_weight_quantizers.232 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.down_proj_weight_quantizers.233 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.down_proj_weight_quantizers.234 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.down_proj_weight_quantizers.235 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.38e-01 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.down_proj_weight_quantizers.236 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.down_proj_weight_quantizers.237 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.down_proj_weight_quantizers.238 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.down_proj_weight_quantizers.239 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.51e-01 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.down_proj_weight_quantizers.240 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.down_proj_weight_quantizers.241 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.down_proj_weight_quantizers.242 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.down_proj_weight_quantizers.243 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.down_proj_weight_quantizers.244 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.down_proj_weight_quantizers.245 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.14e-01 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.down_proj_weight_quantizers.246 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.down_proj_weight_quantizers.247 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.down_proj_weight_quantizers.248 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.down_proj_weight_quantizers.249 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.down_proj_weight_quantizers.250 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.36e-01 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.down_proj_weight_quantizers.251 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.95e-01 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.down_proj_weight_quantizers.252 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.down_proj_weight_quantizers.253 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.down_proj_weight_quantizers.254 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.down_proj_weight_quantizers.255 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.59e-02 calibrator=MaxCalibrator quant) +model.layers.50.mlp.shared_experts.gate_proj.input_quantizer TensorQuantizer(disabled) +model.layers.50.mlp.shared_experts.gate_proj.output_quantizer TensorQuantizer(disabled) +model.layers.50.mlp.shared_experts.gate_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.50.mlp.shared_experts.up_proj.input_quantizer TensorQuantizer(disabled) +model.layers.50.mlp.shared_experts.up_proj.output_quantizer TensorQuantizer(disabled) +model.layers.50.mlp.shared_experts.up_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.50.mlp.shared_experts.down_proj.input_quantizer TensorQuantizer(disabled) +model.layers.50.mlp.shared_experts.down_proj.output_quantizer TensorQuantizer(disabled) +model.layers.50.mlp.shared_experts.down_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.51.self_attn.q_a_proj.input_quantizer TensorQuantizer(disabled) +model.layers.51.self_attn.q_a_proj.output_quantizer TensorQuantizer(disabled) +model.layers.51.self_attn.q_a_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.51.self_attn.q_b_proj.input_quantizer TensorQuantizer(disabled) +model.layers.51.self_attn.q_b_proj.output_quantizer TensorQuantizer(disabled) +model.layers.51.self_attn.q_b_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.51.self_attn.kv_a_proj_with_mqa.input_quantizer TensorQuantizer(disabled) +model.layers.51.self_attn.kv_a_proj_with_mqa.output_quantizer TensorQuantizer(disabled) +model.layers.51.self_attn.kv_a_proj_with_mqa.weight_quantizer TensorQuantizer(disabled) +model.layers.51.self_attn.kv_b_proj.input_quantizer TensorQuantizer(disabled) +model.layers.51.self_attn.kv_b_proj.output_quantizer TensorQuantizer(disabled) +model.layers.51.self_attn.kv_b_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.51.self_attn.o_proj.input_quantizer TensorQuantizer(disabled) +model.layers.51.self_attn.o_proj.output_quantizer TensorQuantizer(disabled) +model.layers.51.self_attn.o_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.51.self_attn.q_bmm_quantizer TensorQuantizer(disabled) +model.layers.51.self_attn.k_bmm_quantizer TensorQuantizer((4, 3) bit fake per-tensor amax=4.48e+02(const) calibrator=MaxCalibrator quant) +model.layers.51.self_attn.v_bmm_quantizer TensorQuantizer((4, 3) bit fake per-tensor amax=4.48e+02(const) calibrator=MaxCalibrator quant) +model.layers.51.self_attn.p_bmm_quantizer TensorQuantizer(disabled) +model.layers.51.mlp.experts.gate_up_proj_input_quantizer TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.62e+00 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.gate_up_proj_weight_quantizers.0 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.gate_up_proj_weight_quantizers.1 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.gate_up_proj_weight_quantizers.2 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.gate_up_proj_weight_quantizers.3 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.gate_up_proj_weight_quantizers.4 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.73e-01 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.gate_up_proj_weight_quantizers.5 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.62e-01 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.gate_up_proj_weight_quantizers.6 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.gate_up_proj_weight_quantizers.7 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.gate_up_proj_weight_quantizers.8 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.gate_up_proj_weight_quantizers.9 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.gate_up_proj_weight_quantizers.10 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.gate_up_proj_weight_quantizers.11 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.gate_up_proj_weight_quantizers.12 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.gate_up_proj_weight_quantizers.13 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.gate_up_proj_weight_quantizers.14 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.gate_up_proj_weight_quantizers.15 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.gate_up_proj_weight_quantizers.16 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.gate_up_proj_weight_quantizers.17 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.gate_up_proj_weight_quantizers.18 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.33e-01 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.gate_up_proj_weight_quantizers.19 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.gate_up_proj_weight_quantizers.20 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.gate_up_proj_weight_quantizers.21 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.gate_up_proj_weight_quantizers.22 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.gate_up_proj_weight_quantizers.23 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.22e-01 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.gate_up_proj_weight_quantizers.24 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.38e-01 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.gate_up_proj_weight_quantizers.25 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.gate_up_proj_weight_quantizers.26 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.gate_up_proj_weight_quantizers.27 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.gate_up_proj_weight_quantizers.28 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.gate_up_proj_weight_quantizers.29 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.gate_up_proj_weight_quantizers.30 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.gate_up_proj_weight_quantizers.31 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.gate_up_proj_weight_quantizers.32 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.gate_up_proj_weight_quantizers.33 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.gate_up_proj_weight_quantizers.34 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.gate_up_proj_weight_quantizers.35 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.gate_up_proj_weight_quantizers.36 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.gate_up_proj_weight_quantizers.37 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.gate_up_proj_weight_quantizers.38 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.gate_up_proj_weight_quantizers.39 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.16e-01 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.gate_up_proj_weight_quantizers.40 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.88e-01 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.gate_up_proj_weight_quantizers.41 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.35e-01 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.gate_up_proj_weight_quantizers.42 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.gate_up_proj_weight_quantizers.43 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.gate_up_proj_weight_quantizers.44 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.gate_up_proj_weight_quantizers.45 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.gate_up_proj_weight_quantizers.46 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.gate_up_proj_weight_quantizers.47 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.gate_up_proj_weight_quantizers.48 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.gate_up_proj_weight_quantizers.49 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.gate_up_proj_weight_quantizers.50 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.gate_up_proj_weight_quantizers.51 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.gate_up_proj_weight_quantizers.52 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.25e-01 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.gate_up_proj_weight_quantizers.53 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.gate_up_proj_weight_quantizers.54 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.gate_up_proj_weight_quantizers.55 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.gate_up_proj_weight_quantizers.56 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.gate_up_proj_weight_quantizers.57 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.18e-01 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.gate_up_proj_weight_quantizers.58 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.gate_up_proj_weight_quantizers.59 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.gate_up_proj_weight_quantizers.60 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.gate_up_proj_weight_quantizers.61 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.gate_up_proj_weight_quantizers.62 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.gate_up_proj_weight_quantizers.63 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.48e-01 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.gate_up_proj_weight_quantizers.64 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.17e-01 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.gate_up_proj_weight_quantizers.65 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.gate_up_proj_weight_quantizers.66 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.gate_up_proj_weight_quantizers.67 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.gate_up_proj_weight_quantizers.68 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.27e-01 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.gate_up_proj_weight_quantizers.69 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.77e-01 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.gate_up_proj_weight_quantizers.70 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.gate_up_proj_weight_quantizers.71 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.80e-01 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.gate_up_proj_weight_quantizers.72 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.14e-01 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.gate_up_proj_weight_quantizers.73 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.gate_up_proj_weight_quantizers.74 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.gate_up_proj_weight_quantizers.75 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.gate_up_proj_weight_quantizers.76 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.gate_up_proj_weight_quantizers.77 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.gate_up_proj_weight_quantizers.78 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.35e-01 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.gate_up_proj_weight_quantizers.79 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.gate_up_proj_weight_quantizers.80 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.55e-01 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.gate_up_proj_weight_quantizers.81 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.gate_up_proj_weight_quantizers.82 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.gate_up_proj_weight_quantizers.83 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.22e-01 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.gate_up_proj_weight_quantizers.84 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.gate_up_proj_weight_quantizers.85 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.gate_up_proj_weight_quantizers.86 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.gate_up_proj_weight_quantizers.87 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.gate_up_proj_weight_quantizers.88 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.gate_up_proj_weight_quantizers.89 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.gate_up_proj_weight_quantizers.90 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.gate_up_proj_weight_quantizers.91 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.gate_up_proj_weight_quantizers.92 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.gate_up_proj_weight_quantizers.93 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.15e-01 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.gate_up_proj_weight_quantizers.94 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.gate_up_proj_weight_quantizers.95 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.gate_up_proj_weight_quantizers.96 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.gate_up_proj_weight_quantizers.97 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.gate_up_proj_weight_quantizers.98 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.gate_up_proj_weight_quantizers.99 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.gate_up_proj_weight_quantizers.100 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.gate_up_proj_weight_quantizers.101 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.gate_up_proj_weight_quantizers.102 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.gate_up_proj_weight_quantizers.103 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.17e-01 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.gate_up_proj_weight_quantizers.104 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.20e-01 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.gate_up_proj_weight_quantizers.105 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.gate_up_proj_weight_quantizers.106 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.gate_up_proj_weight_quantizers.107 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.gate_up_proj_weight_quantizers.108 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.38e-01 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.gate_up_proj_weight_quantizers.109 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.17e-01 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.gate_up_proj_weight_quantizers.110 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.17e-01 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.gate_up_proj_weight_quantizers.111 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.gate_up_proj_weight_quantizers.112 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.gate_up_proj_weight_quantizers.113 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.gate_up_proj_weight_quantizers.114 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.gate_up_proj_weight_quantizers.115 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.gate_up_proj_weight_quantizers.116 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.gate_up_proj_weight_quantizers.117 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.gate_up_proj_weight_quantizers.118 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.gate_up_proj_weight_quantizers.119 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.36e-01 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.gate_up_proj_weight_quantizers.120 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.gate_up_proj_weight_quantizers.121 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.gate_up_proj_weight_quantizers.122 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.gate_up_proj_weight_quantizers.123 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.gate_up_proj_weight_quantizers.124 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.gate_up_proj_weight_quantizers.125 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.45e-01 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.gate_up_proj_weight_quantizers.126 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.gate_up_proj_weight_quantizers.127 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.gate_up_proj_weight_quantizers.128 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.11e-01 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.gate_up_proj_weight_quantizers.129 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.76e-01 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.gate_up_proj_weight_quantizers.130 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.gate_up_proj_weight_quantizers.131 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.40e-01 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.gate_up_proj_weight_quantizers.132 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.gate_up_proj_weight_quantizers.133 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.14e-01 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.gate_up_proj_weight_quantizers.134 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.gate_up_proj_weight_quantizers.135 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.27e-01 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.gate_up_proj_weight_quantizers.136 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.gate_up_proj_weight_quantizers.137 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.gate_up_proj_weight_quantizers.138 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.gate_up_proj_weight_quantizers.139 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.gate_up_proj_weight_quantizers.140 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.gate_up_proj_weight_quantizers.141 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.gate_up_proj_weight_quantizers.142 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.34e-01 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.gate_up_proj_weight_quantizers.143 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.24e-01 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.gate_up_proj_weight_quantizers.144 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.gate_up_proj_weight_quantizers.145 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.31e-01 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.gate_up_proj_weight_quantizers.146 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.gate_up_proj_weight_quantizers.147 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.gate_up_proj_weight_quantizers.148 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.gate_up_proj_weight_quantizers.149 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.gate_up_proj_weight_quantizers.150 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.gate_up_proj_weight_quantizers.151 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.gate_up_proj_weight_quantizers.152 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.gate_up_proj_weight_quantizers.153 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.64e-02 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.gate_up_proj_weight_quantizers.154 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.14e-01 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.gate_up_proj_weight_quantizers.155 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.gate_up_proj_weight_quantizers.156 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.19e-01 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.gate_up_proj_weight_quantizers.157 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.gate_up_proj_weight_quantizers.158 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.42e-01 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.gate_up_proj_weight_quantizers.159 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.gate_up_proj_weight_quantizers.160 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.35e-01 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.gate_up_proj_weight_quantizers.161 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.18e-01 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.gate_up_proj_weight_quantizers.162 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.gate_up_proj_weight_quantizers.163 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.gate_up_proj_weight_quantizers.164 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.28e-01 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.gate_up_proj_weight_quantizers.165 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.gate_up_proj_weight_quantizers.166 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.gate_up_proj_weight_quantizers.167 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.68e-01 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.gate_up_proj_weight_quantizers.168 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.34e-01 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.gate_up_proj_weight_quantizers.169 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.32e-01 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.gate_up_proj_weight_quantizers.170 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.gate_up_proj_weight_quantizers.171 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.gate_up_proj_weight_quantizers.172 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.gate_up_proj_weight_quantizers.173 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.gate_up_proj_weight_quantizers.174 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.gate_up_proj_weight_quantizers.175 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.gate_up_proj_weight_quantizers.176 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.gate_up_proj_weight_quantizers.177 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.gate_up_proj_weight_quantizers.178 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.gate_up_proj_weight_quantizers.179 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.gate_up_proj_weight_quantizers.180 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.16e-01 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.gate_up_proj_weight_quantizers.181 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.16e-01 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.gate_up_proj_weight_quantizers.182 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.gate_up_proj_weight_quantizers.183 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.gate_up_proj_weight_quantizers.184 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.69e-01 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.gate_up_proj_weight_quantizers.185 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.22e-01 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.gate_up_proj_weight_quantizers.186 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.gate_up_proj_weight_quantizers.187 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.gate_up_proj_weight_quantizers.188 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.76e-01 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.gate_up_proj_weight_quantizers.189 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.gate_up_proj_weight_quantizers.190 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.gate_up_proj_weight_quantizers.191 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.gate_up_proj_weight_quantizers.192 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.97e-01 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.gate_up_proj_weight_quantizers.193 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.gate_up_proj_weight_quantizers.194 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.gate_up_proj_weight_quantizers.195 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.gate_up_proj_weight_quantizers.196 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.87e-01 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.gate_up_proj_weight_quantizers.197 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.gate_up_proj_weight_quantizers.198 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.20e-01 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.gate_up_proj_weight_quantizers.199 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.gate_up_proj_weight_quantizers.200 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.gate_up_proj_weight_quantizers.201 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.gate_up_proj_weight_quantizers.202 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.gate_up_proj_weight_quantizers.203 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.gate_up_proj_weight_quantizers.204 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.gate_up_proj_weight_quantizers.205 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.gate_up_proj_weight_quantizers.206 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.gate_up_proj_weight_quantizers.207 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.gate_up_proj_weight_quantizers.208 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.41e-01 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.gate_up_proj_weight_quantizers.209 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.gate_up_proj_weight_quantizers.210 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.gate_up_proj_weight_quantizers.211 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.42e-01 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.gate_up_proj_weight_quantizers.212 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.88e-01 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.gate_up_proj_weight_quantizers.213 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.gate_up_proj_weight_quantizers.214 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.gate_up_proj_weight_quantizers.215 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.gate_up_proj_weight_quantizers.216 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.gate_up_proj_weight_quantizers.217 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.gate_up_proj_weight_quantizers.218 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.gate_up_proj_weight_quantizers.219 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.gate_up_proj_weight_quantizers.220 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.gate_up_proj_weight_quantizers.221 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.gate_up_proj_weight_quantizers.222 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.gate_up_proj_weight_quantizers.223 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.gate_up_proj_weight_quantizers.224 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.gate_up_proj_weight_quantizers.225 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.45e-01 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.gate_up_proj_weight_quantizers.226 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.gate_up_proj_weight_quantizers.227 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.49e-01 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.gate_up_proj_weight_quantizers.228 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.gate_up_proj_weight_quantizers.229 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.07e-01 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.gate_up_proj_weight_quantizers.230 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.76e-01 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.gate_up_proj_weight_quantizers.231 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.gate_up_proj_weight_quantizers.232 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.gate_up_proj_weight_quantizers.233 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.gate_up_proj_weight_quantizers.234 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.gate_up_proj_weight_quantizers.235 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.gate_up_proj_weight_quantizers.236 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.gate_up_proj_weight_quantizers.237 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.gate_up_proj_weight_quantizers.238 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.gate_up_proj_weight_quantizers.239 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.gate_up_proj_weight_quantizers.240 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.gate_up_proj_weight_quantizers.241 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.71e-01 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.gate_up_proj_weight_quantizers.242 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.gate_up_proj_weight_quantizers.243 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.gate_up_proj_weight_quantizers.244 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.gate_up_proj_weight_quantizers.245 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.gate_up_proj_weight_quantizers.246 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.26e-01 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.gate_up_proj_weight_quantizers.247 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.gate_up_proj_weight_quantizers.248 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.gate_up_proj_weight_quantizers.249 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.59e-01 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.gate_up_proj_weight_quantizers.250 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.31e-01 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.gate_up_proj_weight_quantizers.251 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.25e-01 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.gate_up_proj_weight_quantizers.252 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.gate_up_proj_weight_quantizers.253 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.gate_up_proj_weight_quantizers.254 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.gate_up_proj_weight_quantizers.255 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.down_proj_input_quantizer TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.38e+01 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.down_proj_weight_quantizers.0 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.down_proj_weight_quantizers.1 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.down_proj_weight_quantizers.2 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.down_proj_weight_quantizers.3 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.down_proj_weight_quantizers.4 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.81e-01 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.down_proj_weight_quantizers.5 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.02e-01 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.down_proj_weight_quantizers.6 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.down_proj_weight_quantizers.7 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.down_proj_weight_quantizers.8 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.down_proj_weight_quantizers.9 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.40e-01 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.down_proj_weight_quantizers.10 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.down_proj_weight_quantizers.11 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.down_proj_weight_quantizers.12 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.down_proj_weight_quantizers.13 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.down_proj_weight_quantizers.14 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.45e-02 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.down_proj_weight_quantizers.15 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.down_proj_weight_quantizers.16 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.down_proj_weight_quantizers.17 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.down_proj_weight_quantizers.18 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.down_proj_weight_quantizers.19 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.down_proj_weight_quantizers.20 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.down_proj_weight_quantizers.21 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.down_proj_weight_quantizers.22 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.down_proj_weight_quantizers.23 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.down_proj_weight_quantizers.24 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.87e-01 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.down_proj_weight_quantizers.25 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.down_proj_weight_quantizers.26 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.down_proj_weight_quantizers.27 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.40e-02 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.down_proj_weight_quantizers.28 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.15e-01 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.down_proj_weight_quantizers.29 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.50e-02 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.down_proj_weight_quantizers.30 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.down_proj_weight_quantizers.31 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.down_proj_weight_quantizers.32 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.down_proj_weight_quantizers.33 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.down_proj_weight_quantizers.34 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.down_proj_weight_quantizers.35 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.59e-02 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.down_proj_weight_quantizers.36 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.down_proj_weight_quantizers.37 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.down_proj_weight_quantizers.38 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.down_proj_weight_quantizers.39 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.down_proj_weight_quantizers.40 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.down_proj_weight_quantizers.41 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.down_proj_weight_quantizers.42 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.down_proj_weight_quantizers.43 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.down_proj_weight_quantizers.44 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.49e-01 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.down_proj_weight_quantizers.45 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.down_proj_weight_quantizers.46 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.down_proj_weight_quantizers.47 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.17e-01 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.down_proj_weight_quantizers.48 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.down_proj_weight_quantizers.49 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.down_proj_weight_quantizers.50 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.down_proj_weight_quantizers.51 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.down_proj_weight_quantizers.52 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.down_proj_weight_quantizers.53 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.down_proj_weight_quantizers.54 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.down_proj_weight_quantizers.55 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.down_proj_weight_quantizers.56 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.29e-01 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.down_proj_weight_quantizers.57 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.down_proj_weight_quantizers.58 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.19e-01 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.down_proj_weight_quantizers.59 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.down_proj_weight_quantizers.60 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.down_proj_weight_quantizers.61 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.down_proj_weight_quantizers.62 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.down_proj_weight_quantizers.63 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.45e-01 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.down_proj_weight_quantizers.64 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.down_proj_weight_quantizers.65 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.down_proj_weight_quantizers.66 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.24e-01 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.down_proj_weight_quantizers.67 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.down_proj_weight_quantizers.68 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.down_proj_weight_quantizers.69 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.30e-01 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.down_proj_weight_quantizers.70 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.down_proj_weight_quantizers.71 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.down_proj_weight_quantizers.72 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.down_proj_weight_quantizers.73 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.down_proj_weight_quantizers.74 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.59e-02 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.down_proj_weight_quantizers.75 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.down_proj_weight_quantizers.76 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.down_proj_weight_quantizers.77 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.down_proj_weight_quantizers.78 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.92e-01 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.down_proj_weight_quantizers.79 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.54e-02 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.down_proj_weight_quantizers.80 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.25e-01 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.down_proj_weight_quantizers.81 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.down_proj_weight_quantizers.82 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.down_proj_weight_quantizers.83 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.20e-01 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.down_proj_weight_quantizers.84 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.64e-02 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.down_proj_weight_quantizers.85 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.45e-02 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.down_proj_weight_quantizers.86 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.59e-02 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.down_proj_weight_quantizers.87 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.59e-02 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.down_proj_weight_quantizers.88 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.down_proj_weight_quantizers.89 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.down_proj_weight_quantizers.90 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.down_proj_weight_quantizers.91 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.down_proj_weight_quantizers.92 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.down_proj_weight_quantizers.93 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.down_proj_weight_quantizers.94 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.down_proj_weight_quantizers.95 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.down_proj_weight_quantizers.96 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.down_proj_weight_quantizers.97 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.down_proj_weight_quantizers.98 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.down_proj_weight_quantizers.99 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.down_proj_weight_quantizers.100 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.46e-01 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.down_proj_weight_quantizers.101 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.down_proj_weight_quantizers.102 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.down_proj_weight_quantizers.103 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.down_proj_weight_quantizers.104 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.down_proj_weight_quantizers.105 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.down_proj_weight_quantizers.106 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.down_proj_weight_quantizers.107 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.down_proj_weight_quantizers.108 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.down_proj_weight_quantizers.109 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.down_proj_weight_quantizers.110 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.39e-01 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.down_proj_weight_quantizers.111 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.down_proj_weight_quantizers.112 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.down_proj_weight_quantizers.113 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.down_proj_weight_quantizers.114 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.down_proj_weight_quantizers.115 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.down_proj_weight_quantizers.116 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.45e-02 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.down_proj_weight_quantizers.117 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.down_proj_weight_quantizers.118 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.down_proj_weight_quantizers.119 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.30e-01 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.down_proj_weight_quantizers.120 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.down_proj_weight_quantizers.121 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.down_proj_weight_quantizers.122 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.down_proj_weight_quantizers.123 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.down_proj_weight_quantizers.124 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.down_proj_weight_quantizers.125 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.33e-01 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.down_proj_weight_quantizers.126 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.down_proj_weight_quantizers.127 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.down_proj_weight_quantizers.128 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.40e-01 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.down_proj_weight_quantizers.129 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.down_proj_weight_quantizers.130 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.down_proj_weight_quantizers.131 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.46e-01 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.down_proj_weight_quantizers.132 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.down_proj_weight_quantizers.133 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.down_proj_weight_quantizers.134 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.down_proj_weight_quantizers.135 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.50e-02 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.down_proj_weight_quantizers.136 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.down_proj_weight_quantizers.137 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.59e-02 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.down_proj_weight_quantizers.138 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.down_proj_weight_quantizers.139 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.down_proj_weight_quantizers.140 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.down_proj_weight_quantizers.141 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.50e-02 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.down_proj_weight_quantizers.142 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.54e-02 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.down_proj_weight_quantizers.143 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.32e-01 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.down_proj_weight_quantizers.144 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.down_proj_weight_quantizers.145 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.down_proj_weight_quantizers.146 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.down_proj_weight_quantizers.147 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.down_proj_weight_quantizers.148 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.down_proj_weight_quantizers.149 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.17e-01 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.down_proj_weight_quantizers.150 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.down_proj_weight_quantizers.151 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.down_proj_weight_quantizers.152 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.50e-02 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.down_proj_weight_quantizers.153 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.15e-02 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.down_proj_weight_quantizers.154 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.down_proj_weight_quantizers.155 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.down_proj_weight_quantizers.156 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.45e-02 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.down_proj_weight_quantizers.157 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.down_proj_weight_quantizers.158 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.06e-01 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.down_proj_weight_quantizers.159 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.50e-02 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.down_proj_weight_quantizers.160 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.85e-01 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.down_proj_weight_quantizers.161 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.54e-01 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.down_proj_weight_quantizers.162 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.64e-02 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.down_proj_weight_quantizers.163 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.down_proj_weight_quantizers.164 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.down_proj_weight_quantizers.165 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.down_proj_weight_quantizers.166 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.down_proj_weight_quantizers.167 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.33e-01 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.down_proj_weight_quantizers.168 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.down_proj_weight_quantizers.169 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.down_proj_weight_quantizers.170 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.down_proj_weight_quantizers.171 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.down_proj_weight_quantizers.172 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.down_proj_weight_quantizers.173 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.down_proj_weight_quantizers.174 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.down_proj_weight_quantizers.175 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.down_proj_weight_quantizers.176 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.down_proj_weight_quantizers.177 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.down_proj_weight_quantizers.178 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.20e-01 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.down_proj_weight_quantizers.179 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.59e-02 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.down_proj_weight_quantizers.180 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.down_proj_weight_quantizers.181 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.down_proj_weight_quantizers.182 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.down_proj_weight_quantizers.183 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.40e-02 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.down_proj_weight_quantizers.184 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.25e-01 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.down_proj_weight_quantizers.185 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.down_proj_weight_quantizers.186 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.down_proj_weight_quantizers.187 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.down_proj_weight_quantizers.188 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.down_proj_weight_quantizers.189 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.down_proj_weight_quantizers.190 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.down_proj_weight_quantizers.191 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.down_proj_weight_quantizers.192 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.down_proj_weight_quantizers.193 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.down_proj_weight_quantizers.194 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.down_proj_weight_quantizers.195 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.down_proj_weight_quantizers.196 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.05e-01 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.down_proj_weight_quantizers.197 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.down_proj_weight_quantizers.198 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.down_proj_weight_quantizers.199 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.down_proj_weight_quantizers.200 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.54e-02 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.down_proj_weight_quantizers.201 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.down_proj_weight_quantizers.202 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.down_proj_weight_quantizers.203 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.down_proj_weight_quantizers.204 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.54e-02 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.down_proj_weight_quantizers.205 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.down_proj_weight_quantizers.206 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.54e-02 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.down_proj_weight_quantizers.207 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.down_proj_weight_quantizers.208 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.56e-01 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.down_proj_weight_quantizers.209 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.down_proj_weight_quantizers.210 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.down_proj_weight_quantizers.211 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.27e-01 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.down_proj_weight_quantizers.212 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.24e-01 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.down_proj_weight_quantizers.213 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.down_proj_weight_quantizers.214 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.down_proj_weight_quantizers.215 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.down_proj_weight_quantizers.216 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.down_proj_weight_quantizers.217 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.down_proj_weight_quantizers.218 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.down_proj_weight_quantizers.219 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.down_proj_weight_quantizers.220 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.down_proj_weight_quantizers.221 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.80e-01 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.down_proj_weight_quantizers.222 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.down_proj_weight_quantizers.223 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.down_proj_weight_quantizers.224 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.down_proj_weight_quantizers.225 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.down_proj_weight_quantizers.226 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.down_proj_weight_quantizers.227 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.46e-01 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.down_proj_weight_quantizers.228 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.down_proj_weight_quantizers.229 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.99e-01 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.down_proj_weight_quantizers.230 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.down_proj_weight_quantizers.231 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.64e-02 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.down_proj_weight_quantizers.232 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.down_proj_weight_quantizers.233 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.down_proj_weight_quantizers.234 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.down_proj_weight_quantizers.235 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.down_proj_weight_quantizers.236 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.down_proj_weight_quantizers.237 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.down_proj_weight_quantizers.238 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.down_proj_weight_quantizers.239 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.down_proj_weight_quantizers.240 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.down_proj_weight_quantizers.241 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.25e-01 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.down_proj_weight_quantizers.242 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.down_proj_weight_quantizers.243 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.down_proj_weight_quantizers.244 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.down_proj_weight_quantizers.245 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.down_proj_weight_quantizers.246 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.78e-01 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.down_proj_weight_quantizers.247 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.down_proj_weight_quantizers.248 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.down_proj_weight_quantizers.249 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.19e-01 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.down_proj_weight_quantizers.250 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.down_proj_weight_quantizers.251 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.down_proj_weight_quantizers.252 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.down_proj_weight_quantizers.253 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.down_proj_weight_quantizers.254 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.down_proj_weight_quantizers.255 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.51.mlp.shared_experts.gate_proj.input_quantizer TensorQuantizer(disabled) +model.layers.51.mlp.shared_experts.gate_proj.output_quantizer TensorQuantizer(disabled) +model.layers.51.mlp.shared_experts.gate_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.51.mlp.shared_experts.up_proj.input_quantizer TensorQuantizer(disabled) +model.layers.51.mlp.shared_experts.up_proj.output_quantizer TensorQuantizer(disabled) +model.layers.51.mlp.shared_experts.up_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.51.mlp.shared_experts.down_proj.input_quantizer TensorQuantizer(disabled) +model.layers.51.mlp.shared_experts.down_proj.output_quantizer TensorQuantizer(disabled) +model.layers.51.mlp.shared_experts.down_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.52.self_attn.q_a_proj.input_quantizer TensorQuantizer(disabled) +model.layers.52.self_attn.q_a_proj.output_quantizer TensorQuantizer(disabled) +model.layers.52.self_attn.q_a_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.52.self_attn.q_b_proj.input_quantizer TensorQuantizer(disabled) +model.layers.52.self_attn.q_b_proj.output_quantizer TensorQuantizer(disabled) +model.layers.52.self_attn.q_b_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.52.self_attn.kv_a_proj_with_mqa.input_quantizer TensorQuantizer(disabled) +model.layers.52.self_attn.kv_a_proj_with_mqa.output_quantizer TensorQuantizer(disabled) +model.layers.52.self_attn.kv_a_proj_with_mqa.weight_quantizer TensorQuantizer(disabled) +model.layers.52.self_attn.kv_b_proj.input_quantizer TensorQuantizer(disabled) +model.layers.52.self_attn.kv_b_proj.output_quantizer TensorQuantizer(disabled) +model.layers.52.self_attn.kv_b_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.52.self_attn.o_proj.input_quantizer TensorQuantizer(disabled) +model.layers.52.self_attn.o_proj.output_quantizer TensorQuantizer(disabled) +model.layers.52.self_attn.o_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.52.self_attn.q_bmm_quantizer TensorQuantizer(disabled) +model.layers.52.self_attn.k_bmm_quantizer TensorQuantizer((4, 3) bit fake per-tensor amax=4.48e+02(const) calibrator=MaxCalibrator quant) +model.layers.52.self_attn.v_bmm_quantizer TensorQuantizer((4, 3) bit fake per-tensor amax=4.48e+02(const) calibrator=MaxCalibrator quant) +model.layers.52.self_attn.p_bmm_quantizer TensorQuantizer(disabled) +model.layers.52.mlp.experts.gate_up_proj_input_quantizer TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.80e+00 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.gate_up_proj_weight_quantizers.0 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.gate_up_proj_weight_quantizers.1 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.gate_up_proj_weight_quantizers.2 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.gate_up_proj_weight_quantizers.3 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.gate_up_proj_weight_quantizers.4 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.gate_up_proj_weight_quantizers.5 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.gate_up_proj_weight_quantizers.6 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.gate_up_proj_weight_quantizers.7 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.72e-01 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.gate_up_proj_weight_quantizers.8 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.gate_up_proj_weight_quantizers.9 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.47e-01 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.gate_up_proj_weight_quantizers.10 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.gate_up_proj_weight_quantizers.11 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.gate_up_proj_weight_quantizers.12 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.gate_up_proj_weight_quantizers.13 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.18e-01 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.gate_up_proj_weight_quantizers.14 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.gate_up_proj_weight_quantizers.15 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.gate_up_proj_weight_quantizers.16 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.gate_up_proj_weight_quantizers.17 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.54e-01 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.gate_up_proj_weight_quantizers.18 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.17e-01 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.gate_up_proj_weight_quantizers.19 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.90e-01 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.gate_up_proj_weight_quantizers.20 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.gate_up_proj_weight_quantizers.21 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.gate_up_proj_weight_quantizers.22 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.gate_up_proj_weight_quantizers.23 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.79e-01 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.gate_up_proj_weight_quantizers.24 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.gate_up_proj_weight_quantizers.25 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.gate_up_proj_weight_quantizers.26 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.10e-01 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.gate_up_proj_weight_quantizers.27 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.gate_up_proj_weight_quantizers.28 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.gate_up_proj_weight_quantizers.29 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.18e-01 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.gate_up_proj_weight_quantizers.30 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.gate_up_proj_weight_quantizers.31 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.gate_up_proj_weight_quantizers.32 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.gate_up_proj_weight_quantizers.33 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.gate_up_proj_weight_quantizers.34 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.gate_up_proj_weight_quantizers.35 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.gate_up_proj_weight_quantizers.36 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.gate_up_proj_weight_quantizers.37 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.gate_up_proj_weight_quantizers.38 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.gate_up_proj_weight_quantizers.39 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.gate_up_proj_weight_quantizers.40 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.gate_up_proj_weight_quantizers.41 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.gate_up_proj_weight_quantizers.42 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.gate_up_proj_weight_quantizers.43 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.gate_up_proj_weight_quantizers.44 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.gate_up_proj_weight_quantizers.45 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.gate_up_proj_weight_quantizers.46 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.gate_up_proj_weight_quantizers.47 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.gate_up_proj_weight_quantizers.48 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.gate_up_proj_weight_quantizers.49 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.gate_up_proj_weight_quantizers.50 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.gate_up_proj_weight_quantizers.51 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.46e-01 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.gate_up_proj_weight_quantizers.52 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.gate_up_proj_weight_quantizers.53 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.gate_up_proj_weight_quantizers.54 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.gate_up_proj_weight_quantizers.55 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.gate_up_proj_weight_quantizers.56 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.gate_up_proj_weight_quantizers.57 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.gate_up_proj_weight_quantizers.58 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.91e-01 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.gate_up_proj_weight_quantizers.59 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.74e-01 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.gate_up_proj_weight_quantizers.60 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.gate_up_proj_weight_quantizers.61 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.gate_up_proj_weight_quantizers.62 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.gate_up_proj_weight_quantizers.63 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.gate_up_proj_weight_quantizers.64 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.gate_up_proj_weight_quantizers.65 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.gate_up_proj_weight_quantizers.66 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.gate_up_proj_weight_quantizers.67 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.gate_up_proj_weight_quantizers.68 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.gate_up_proj_weight_quantizers.69 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.48e-01 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.gate_up_proj_weight_quantizers.70 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.57e-01 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.gate_up_proj_weight_quantizers.71 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.gate_up_proj_weight_quantizers.72 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.gate_up_proj_weight_quantizers.73 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.gate_up_proj_weight_quantizers.74 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.gate_up_proj_weight_quantizers.75 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.gate_up_proj_weight_quantizers.76 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.14e-01 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.gate_up_proj_weight_quantizers.77 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.gate_up_proj_weight_quantizers.78 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.gate_up_proj_weight_quantizers.79 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.97e-01 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.gate_up_proj_weight_quantizers.80 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.gate_up_proj_weight_quantizers.81 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.gate_up_proj_weight_quantizers.82 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.gate_up_proj_weight_quantizers.83 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.14e-01 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.gate_up_proj_weight_quantizers.84 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.gate_up_proj_weight_quantizers.85 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.39e-01 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.gate_up_proj_weight_quantizers.86 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.65e-01 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.gate_up_proj_weight_quantizers.87 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.gate_up_proj_weight_quantizers.88 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.gate_up_proj_weight_quantizers.89 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.gate_up_proj_weight_quantizers.90 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.gate_up_proj_weight_quantizers.91 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.26e-01 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.gate_up_proj_weight_quantizers.92 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.gate_up_proj_weight_quantizers.93 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.gate_up_proj_weight_quantizers.94 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.21e-01 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.gate_up_proj_weight_quantizers.95 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.gate_up_proj_weight_quantizers.96 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.gate_up_proj_weight_quantizers.97 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.gate_up_proj_weight_quantizers.98 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.gate_up_proj_weight_quantizers.99 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.gate_up_proj_weight_quantizers.100 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.87e-01 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.gate_up_proj_weight_quantizers.101 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.gate_up_proj_weight_quantizers.102 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.gate_up_proj_weight_quantizers.103 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.21e-01 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.gate_up_proj_weight_quantizers.104 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.27e-01 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.gate_up_proj_weight_quantizers.105 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.gate_up_proj_weight_quantizers.106 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.gate_up_proj_weight_quantizers.107 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.gate_up_proj_weight_quantizers.108 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.gate_up_proj_weight_quantizers.109 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.54e-01 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.gate_up_proj_weight_quantizers.110 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.45e-01 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.gate_up_proj_weight_quantizers.111 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.gate_up_proj_weight_quantizers.112 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.gate_up_proj_weight_quantizers.113 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.41e-01 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.gate_up_proj_weight_quantizers.114 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.gate_up_proj_weight_quantizers.115 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.gate_up_proj_weight_quantizers.116 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.gate_up_proj_weight_quantizers.117 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.08e-01 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.gate_up_proj_weight_quantizers.118 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.76e-01 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.gate_up_proj_weight_quantizers.119 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.gate_up_proj_weight_quantizers.120 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.83e-01 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.gate_up_proj_weight_quantizers.121 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.gate_up_proj_weight_quantizers.122 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.gate_up_proj_weight_quantizers.123 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.gate_up_proj_weight_quantizers.124 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.10e-01 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.gate_up_proj_weight_quantizers.125 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.gate_up_proj_weight_quantizers.126 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.gate_up_proj_weight_quantizers.127 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.gate_up_proj_weight_quantizers.128 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.gate_up_proj_weight_quantizers.129 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.gate_up_proj_weight_quantizers.130 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.gate_up_proj_weight_quantizers.131 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.53e-01 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.gate_up_proj_weight_quantizers.132 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.gate_up_proj_weight_quantizers.133 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.gate_up_proj_weight_quantizers.134 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.gate_up_proj_weight_quantizers.135 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.gate_up_proj_weight_quantizers.136 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.gate_up_proj_weight_quantizers.137 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.gate_up_proj_weight_quantizers.138 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.43e-01 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.gate_up_proj_weight_quantizers.139 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.gate_up_proj_weight_quantizers.140 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.28e-01 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.gate_up_proj_weight_quantizers.141 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.gate_up_proj_weight_quantizers.142 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.09e-01 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.gate_up_proj_weight_quantizers.143 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.53e-01 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.gate_up_proj_weight_quantizers.144 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.47e-01 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.gate_up_proj_weight_quantizers.145 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.gate_up_proj_weight_quantizers.146 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.gate_up_proj_weight_quantizers.147 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.gate_up_proj_weight_quantizers.148 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.gate_up_proj_weight_quantizers.149 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.gate_up_proj_weight_quantizers.150 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.32e-01 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.gate_up_proj_weight_quantizers.151 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.25e-01 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.gate_up_proj_weight_quantizers.152 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.gate_up_proj_weight_quantizers.153 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.gate_up_proj_weight_quantizers.154 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.gate_up_proj_weight_quantizers.155 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.gate_up_proj_weight_quantizers.156 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.13e-01 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.gate_up_proj_weight_quantizers.157 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.gate_up_proj_weight_quantizers.158 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.gate_up_proj_weight_quantizers.159 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.gate_up_proj_weight_quantizers.160 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.gate_up_proj_weight_quantizers.161 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.gate_up_proj_weight_quantizers.162 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.gate_up_proj_weight_quantizers.163 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.15e-01 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.gate_up_proj_weight_quantizers.164 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.gate_up_proj_weight_quantizers.165 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.15e-01 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.gate_up_proj_weight_quantizers.166 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.gate_up_proj_weight_quantizers.167 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.gate_up_proj_weight_quantizers.168 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.gate_up_proj_weight_quantizers.169 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.gate_up_proj_weight_quantizers.170 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.gate_up_proj_weight_quantizers.171 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.gate_up_proj_weight_quantizers.172 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.gate_up_proj_weight_quantizers.173 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.gate_up_proj_weight_quantizers.174 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.17e-01 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.gate_up_proj_weight_quantizers.175 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.gate_up_proj_weight_quantizers.176 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.gate_up_proj_weight_quantizers.177 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.17e-01 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.gate_up_proj_weight_quantizers.178 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.46e-01 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.gate_up_proj_weight_quantizers.179 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.gate_up_proj_weight_quantizers.180 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.gate_up_proj_weight_quantizers.181 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.gate_up_proj_weight_quantizers.182 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.gate_up_proj_weight_quantizers.183 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.gate_up_proj_weight_quantizers.184 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.gate_up_proj_weight_quantizers.185 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.gate_up_proj_weight_quantizers.186 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.gate_up_proj_weight_quantizers.187 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.gate_up_proj_weight_quantizers.188 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.34e-01 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.gate_up_proj_weight_quantizers.189 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.gate_up_proj_weight_quantizers.190 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.gate_up_proj_weight_quantizers.191 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.gate_up_proj_weight_quantizers.192 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.gate_up_proj_weight_quantizers.193 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.gate_up_proj_weight_quantizers.194 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.19e-01 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.gate_up_proj_weight_quantizers.195 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.gate_up_proj_weight_quantizers.196 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.gate_up_proj_weight_quantizers.197 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.gate_up_proj_weight_quantizers.198 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.gate_up_proj_weight_quantizers.199 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.64e-02 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.gate_up_proj_weight_quantizers.200 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.gate_up_proj_weight_quantizers.201 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.gate_up_proj_weight_quantizers.202 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.gate_up_proj_weight_quantizers.203 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.gate_up_proj_weight_quantizers.204 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.gate_up_proj_weight_quantizers.205 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.16e-01 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.gate_up_proj_weight_quantizers.206 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.gate_up_proj_weight_quantizers.207 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.gate_up_proj_weight_quantizers.208 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.gate_up_proj_weight_quantizers.209 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.gate_up_proj_weight_quantizers.210 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.gate_up_proj_weight_quantizers.211 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.gate_up_proj_weight_quantizers.212 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.gate_up_proj_weight_quantizers.213 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.25e-01 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.gate_up_proj_weight_quantizers.214 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.33e-01 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.gate_up_proj_weight_quantizers.215 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.gate_up_proj_weight_quantizers.216 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.gate_up_proj_weight_quantizers.217 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.40e-01 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.gate_up_proj_weight_quantizers.218 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.gate_up_proj_weight_quantizers.219 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.66e-01 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.gate_up_proj_weight_quantizers.220 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.gate_up_proj_weight_quantizers.221 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.28e-01 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.gate_up_proj_weight_quantizers.222 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.gate_up_proj_weight_quantizers.223 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.gate_up_proj_weight_quantizers.224 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.gate_up_proj_weight_quantizers.225 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.gate_up_proj_weight_quantizers.226 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.gate_up_proj_weight_quantizers.227 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.gate_up_proj_weight_quantizers.228 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.22e-01 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.gate_up_proj_weight_quantizers.229 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.gate_up_proj_weight_quantizers.230 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.84e-01 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.gate_up_proj_weight_quantizers.231 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.gate_up_proj_weight_quantizers.232 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.gate_up_proj_weight_quantizers.233 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.gate_up_proj_weight_quantizers.234 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.gate_up_proj_weight_quantizers.235 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.gate_up_proj_weight_quantizers.236 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.gate_up_proj_weight_quantizers.237 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.gate_up_proj_weight_quantizers.238 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.gate_up_proj_weight_quantizers.239 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.18e-01 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.gate_up_proj_weight_quantizers.240 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.gate_up_proj_weight_quantizers.241 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.gate_up_proj_weight_quantizers.242 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.gate_up_proj_weight_quantizers.243 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.gate_up_proj_weight_quantizers.244 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.gate_up_proj_weight_quantizers.245 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.19e-01 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.gate_up_proj_weight_quantizers.246 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.gate_up_proj_weight_quantizers.247 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.gate_up_proj_weight_quantizers.248 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.gate_up_proj_weight_quantizers.249 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.31e-01 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.gate_up_proj_weight_quantizers.250 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.gate_up_proj_weight_quantizers.251 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.gate_up_proj_weight_quantizers.252 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.gate_up_proj_weight_quantizers.253 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.gate_up_proj_weight_quantizers.254 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.gate_up_proj_weight_quantizers.255 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.down_proj_input_quantizer TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.62e+01 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.down_proj_weight_quantizers.0 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.31e-01 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.down_proj_weight_quantizers.1 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.down_proj_weight_quantizers.2 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.down_proj_weight_quantizers.3 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.down_proj_weight_quantizers.4 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.down_proj_weight_quantizers.5 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.down_proj_weight_quantizers.6 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.50e-01 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.down_proj_weight_quantizers.7 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.00e-01 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.down_proj_weight_quantizers.8 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.down_proj_weight_quantizers.9 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.down_proj_weight_quantizers.10 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.down_proj_weight_quantizers.11 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.40e-02 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.down_proj_weight_quantizers.12 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.down_proj_weight_quantizers.13 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.down_proj_weight_quantizers.14 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.down_proj_weight_quantizers.15 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.down_proj_weight_quantizers.16 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.down_proj_weight_quantizers.17 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.73e-01 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.down_proj_weight_quantizers.18 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.down_proj_weight_quantizers.19 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.38e-01 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.down_proj_weight_quantizers.20 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.down_proj_weight_quantizers.21 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.down_proj_weight_quantizers.22 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.down_proj_weight_quantizers.23 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.down_proj_weight_quantizers.24 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.40e-02 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.down_proj_weight_quantizers.25 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.down_proj_weight_quantizers.26 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.down_proj_weight_quantizers.27 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.59e-02 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.down_proj_weight_quantizers.28 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.64e-02 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.down_proj_weight_quantizers.29 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.45e-02 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.down_proj_weight_quantizers.30 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.down_proj_weight_quantizers.31 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.down_proj_weight_quantizers.32 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.down_proj_weight_quantizers.33 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.down_proj_weight_quantizers.34 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.down_proj_weight_quantizers.35 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.down_proj_weight_quantizers.36 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.down_proj_weight_quantizers.37 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.down_proj_weight_quantizers.38 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.down_proj_weight_quantizers.39 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.down_proj_weight_quantizers.40 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.down_proj_weight_quantizers.41 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.down_proj_weight_quantizers.42 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.down_proj_weight_quantizers.43 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.down_proj_weight_quantizers.44 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.down_proj_weight_quantizers.45 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.down_proj_weight_quantizers.46 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.down_proj_weight_quantizers.47 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.down_proj_weight_quantizers.48 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.down_proj_weight_quantizers.49 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.down_proj_weight_quantizers.50 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.down_proj_weight_quantizers.51 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.down_proj_weight_quantizers.52 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.25e-01 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.down_proj_weight_quantizers.53 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.down_proj_weight_quantizers.54 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.down_proj_weight_quantizers.55 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.down_proj_weight_quantizers.56 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.down_proj_weight_quantizers.57 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.down_proj_weight_quantizers.58 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.down_proj_weight_quantizers.59 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.20e-01 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.down_proj_weight_quantizers.60 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.64e-02 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.down_proj_weight_quantizers.61 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.down_proj_weight_quantizers.62 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.down_proj_weight_quantizers.63 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.down_proj_weight_quantizers.64 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.down_proj_weight_quantizers.65 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.down_proj_weight_quantizers.66 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.down_proj_weight_quantizers.67 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.down_proj_weight_quantizers.68 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.down_proj_weight_quantizers.69 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.down_proj_weight_quantizers.70 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.down_proj_weight_quantizers.71 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.down_proj_weight_quantizers.72 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.30e-02 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.down_proj_weight_quantizers.73 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.down_proj_weight_quantizers.74 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.down_proj_weight_quantizers.75 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.down_proj_weight_quantizers.76 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.down_proj_weight_quantizers.77 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.down_proj_weight_quantizers.78 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.down_proj_weight_quantizers.79 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.down_proj_weight_quantizers.80 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.down_proj_weight_quantizers.81 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.down_proj_weight_quantizers.82 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.79e-01 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.down_proj_weight_quantizers.83 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.08e-01 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.down_proj_weight_quantizers.84 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.54e-02 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.down_proj_weight_quantizers.85 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.down_proj_weight_quantizers.86 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.down_proj_weight_quantizers.87 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.down_proj_weight_quantizers.88 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.down_proj_weight_quantizers.89 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.54e-02 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.down_proj_weight_quantizers.90 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.59e-02 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.down_proj_weight_quantizers.91 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.down_proj_weight_quantizers.92 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.down_proj_weight_quantizers.93 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.25e-01 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.down_proj_weight_quantizers.94 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.down_proj_weight_quantizers.95 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.down_proj_weight_quantizers.96 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.down_proj_weight_quantizers.97 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.down_proj_weight_quantizers.98 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.down_proj_weight_quantizers.99 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.down_proj_weight_quantizers.100 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.down_proj_weight_quantizers.101 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.down_proj_weight_quantizers.102 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.down_proj_weight_quantizers.103 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.down_proj_weight_quantizers.104 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.down_proj_weight_quantizers.105 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.down_proj_weight_quantizers.106 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.down_proj_weight_quantizers.107 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.97e-01 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.down_proj_weight_quantizers.108 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.down_proj_weight_quantizers.109 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.down_proj_weight_quantizers.110 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.down_proj_weight_quantizers.111 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.down_proj_weight_quantizers.112 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.down_proj_weight_quantizers.113 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.37e-01 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.down_proj_weight_quantizers.114 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.down_proj_weight_quantizers.115 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.down_proj_weight_quantizers.116 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.down_proj_weight_quantizers.117 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.down_proj_weight_quantizers.118 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.down_proj_weight_quantizers.119 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.down_proj_weight_quantizers.120 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.46e-01 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.down_proj_weight_quantizers.121 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.down_proj_weight_quantizers.122 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.48e-01 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.down_proj_weight_quantizers.123 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.down_proj_weight_quantizers.124 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.33e-01 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.down_proj_weight_quantizers.125 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.down_proj_weight_quantizers.126 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.down_proj_weight_quantizers.127 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.down_proj_weight_quantizers.128 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.down_proj_weight_quantizers.129 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.down_proj_weight_quantizers.130 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.down_proj_weight_quantizers.131 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.22e-01 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.down_proj_weight_quantizers.132 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.down_proj_weight_quantizers.133 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.50e-02 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.down_proj_weight_quantizers.134 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.down_proj_weight_quantizers.135 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.down_proj_weight_quantizers.136 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.down_proj_weight_quantizers.137 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.down_proj_weight_quantizers.138 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.down_proj_weight_quantizers.139 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.down_proj_weight_quantizers.140 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.down_proj_weight_quantizers.141 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.down_proj_weight_quantizers.142 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.32e-01 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.down_proj_weight_quantizers.143 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.90e-01 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.down_proj_weight_quantizers.144 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.down_proj_weight_quantizers.145 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.down_proj_weight_quantizers.146 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.down_proj_weight_quantizers.147 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.59e-02 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.down_proj_weight_quantizers.148 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.down_proj_weight_quantizers.149 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.down_proj_weight_quantizers.150 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.down_proj_weight_quantizers.151 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.15e-01 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.down_proj_weight_quantizers.152 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.down_proj_weight_quantizers.153 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.down_proj_weight_quantizers.154 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.50e-02 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.down_proj_weight_quantizers.155 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.down_proj_weight_quantizers.156 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.down_proj_weight_quantizers.157 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.down_proj_weight_quantizers.158 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.down_proj_weight_quantizers.159 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.down_proj_weight_quantizers.160 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.down_proj_weight_quantizers.161 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.90e-01 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.down_proj_weight_quantizers.162 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.down_proj_weight_quantizers.163 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.down_proj_weight_quantizers.164 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.down_proj_weight_quantizers.165 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.down_proj_weight_quantizers.166 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.down_proj_weight_quantizers.167 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.down_proj_weight_quantizers.168 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.down_proj_weight_quantizers.169 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.down_proj_weight_quantizers.170 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.down_proj_weight_quantizers.171 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.down_proj_weight_quantizers.172 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.64e-02 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.down_proj_weight_quantizers.173 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.down_proj_weight_quantizers.174 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.46e-01 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.down_proj_weight_quantizers.175 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.down_proj_weight_quantizers.176 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.down_proj_weight_quantizers.177 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.down_proj_weight_quantizers.178 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.down_proj_weight_quantizers.179 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.down_proj_weight_quantizers.180 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.down_proj_weight_quantizers.181 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.down_proj_weight_quantizers.182 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.down_proj_weight_quantizers.183 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.down_proj_weight_quantizers.184 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.45e-02 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.down_proj_weight_quantizers.185 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.down_proj_weight_quantizers.186 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.22e-01 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.down_proj_weight_quantizers.187 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.down_proj_weight_quantizers.188 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.down_proj_weight_quantizers.189 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.down_proj_weight_quantizers.190 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.down_proj_weight_quantizers.191 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.down_proj_weight_quantizers.192 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.down_proj_weight_quantizers.193 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.39e-01 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.down_proj_weight_quantizers.194 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.down_proj_weight_quantizers.195 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.down_proj_weight_quantizers.196 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.down_proj_weight_quantizers.197 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.down_proj_weight_quantizers.198 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.down_proj_weight_quantizers.199 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.down_proj_weight_quantizers.200 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.down_proj_weight_quantizers.201 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.down_proj_weight_quantizers.202 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.down_proj_weight_quantizers.203 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.down_proj_weight_quantizers.204 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.down_proj_weight_quantizers.205 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.down_proj_weight_quantizers.206 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.down_proj_weight_quantizers.207 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.down_proj_weight_quantizers.208 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.down_proj_weight_quantizers.209 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.down_proj_weight_quantizers.210 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.down_proj_weight_quantizers.211 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.down_proj_weight_quantizers.212 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.down_proj_weight_quantizers.213 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.down_proj_weight_quantizers.214 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.down_proj_weight_quantizers.215 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.59e-02 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.down_proj_weight_quantizers.216 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.down_proj_weight_quantizers.217 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.49e-01 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.down_proj_weight_quantizers.218 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.30e-01 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.down_proj_weight_quantizers.219 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.21e-01 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.down_proj_weight_quantizers.220 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.down_proj_weight_quantizers.221 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.down_proj_weight_quantizers.222 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.down_proj_weight_quantizers.223 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.down_proj_weight_quantizers.224 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.down_proj_weight_quantizers.225 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.down_proj_weight_quantizers.226 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.down_proj_weight_quantizers.227 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.down_proj_weight_quantizers.228 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.down_proj_weight_quantizers.229 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.down_proj_weight_quantizers.230 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.26e-01 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.down_proj_weight_quantizers.231 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.54e-02 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.down_proj_weight_quantizers.232 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.down_proj_weight_quantizers.233 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.down_proj_weight_quantizers.234 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.down_proj_weight_quantizers.235 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.down_proj_weight_quantizers.236 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.41e-01 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.down_proj_weight_quantizers.237 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.down_proj_weight_quantizers.238 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.down_proj_weight_quantizers.239 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.down_proj_weight_quantizers.240 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.down_proj_weight_quantizers.241 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.down_proj_weight_quantizers.242 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.down_proj_weight_quantizers.243 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.down_proj_weight_quantizers.244 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.down_proj_weight_quantizers.245 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.down_proj_weight_quantizers.246 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.down_proj_weight_quantizers.247 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.down_proj_weight_quantizers.248 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.down_proj_weight_quantizers.249 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.down_proj_weight_quantizers.250 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.down_proj_weight_quantizers.251 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.down_proj_weight_quantizers.252 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.54e-02 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.down_proj_weight_quantizers.253 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.down_proj_weight_quantizers.254 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.down_proj_weight_quantizers.255 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.64e-02 calibrator=MaxCalibrator quant) +model.layers.52.mlp.shared_experts.gate_proj.input_quantizer TensorQuantizer(disabled) +model.layers.52.mlp.shared_experts.gate_proj.output_quantizer TensorQuantizer(disabled) +model.layers.52.mlp.shared_experts.gate_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.52.mlp.shared_experts.up_proj.input_quantizer TensorQuantizer(disabled) +model.layers.52.mlp.shared_experts.up_proj.output_quantizer TensorQuantizer(disabled) +model.layers.52.mlp.shared_experts.up_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.52.mlp.shared_experts.down_proj.input_quantizer TensorQuantizer(disabled) +model.layers.52.mlp.shared_experts.down_proj.output_quantizer TensorQuantizer(disabled) +model.layers.52.mlp.shared_experts.down_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.53.self_attn.q_a_proj.input_quantizer TensorQuantizer(disabled) +model.layers.53.self_attn.q_a_proj.output_quantizer TensorQuantizer(disabled) +model.layers.53.self_attn.q_a_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.53.self_attn.q_b_proj.input_quantizer TensorQuantizer(disabled) +model.layers.53.self_attn.q_b_proj.output_quantizer TensorQuantizer(disabled) +model.layers.53.self_attn.q_b_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.53.self_attn.kv_a_proj_with_mqa.input_quantizer TensorQuantizer(disabled) +model.layers.53.self_attn.kv_a_proj_with_mqa.output_quantizer TensorQuantizer(disabled) +model.layers.53.self_attn.kv_a_proj_with_mqa.weight_quantizer TensorQuantizer(disabled) +model.layers.53.self_attn.kv_b_proj.input_quantizer TensorQuantizer(disabled) +model.layers.53.self_attn.kv_b_proj.output_quantizer TensorQuantizer(disabled) +model.layers.53.self_attn.kv_b_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.53.self_attn.o_proj.input_quantizer TensorQuantizer(disabled) +model.layers.53.self_attn.o_proj.output_quantizer TensorQuantizer(disabled) +model.layers.53.self_attn.o_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.53.self_attn.q_bmm_quantizer TensorQuantizer(disabled) +model.layers.53.self_attn.k_bmm_quantizer TensorQuantizer((4, 3) bit fake per-tensor amax=4.48e+02(const) calibrator=MaxCalibrator quant) +model.layers.53.self_attn.v_bmm_quantizer TensorQuantizer((4, 3) bit fake per-tensor amax=4.48e+02(const) calibrator=MaxCalibrator quant) +model.layers.53.self_attn.p_bmm_quantizer TensorQuantizer(disabled) +model.layers.53.mlp.experts.gate_up_proj_input_quantizer TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.56e+00 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.gate_up_proj_weight_quantizers.0 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.gate_up_proj_weight_quantizers.1 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.15e-01 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.gate_up_proj_weight_quantizers.2 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.gate_up_proj_weight_quantizers.3 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.gate_up_proj_weight_quantizers.4 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.gate_up_proj_weight_quantizers.5 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.29e-01 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.gate_up_proj_weight_quantizers.6 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.gate_up_proj_weight_quantizers.7 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.gate_up_proj_weight_quantizers.8 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.69e-01 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.gate_up_proj_weight_quantizers.9 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.gate_up_proj_weight_quantizers.10 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.gate_up_proj_weight_quantizers.11 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.gate_up_proj_weight_quantizers.12 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.gate_up_proj_weight_quantizers.13 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.gate_up_proj_weight_quantizers.14 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.gate_up_proj_weight_quantizers.15 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.19e-01 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.gate_up_proj_weight_quantizers.16 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.gate_up_proj_weight_quantizers.17 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.gate_up_proj_weight_quantizers.18 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.14e-01 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.gate_up_proj_weight_quantizers.19 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.gate_up_proj_weight_quantizers.20 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.gate_up_proj_weight_quantizers.21 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.gate_up_proj_weight_quantizers.22 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.gate_up_proj_weight_quantizers.23 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.gate_up_proj_weight_quantizers.24 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.gate_up_proj_weight_quantizers.25 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.21e-01 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.gate_up_proj_weight_quantizers.26 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.gate_up_proj_weight_quantizers.27 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.gate_up_proj_weight_quantizers.28 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.gate_up_proj_weight_quantizers.29 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.gate_up_proj_weight_quantizers.30 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.gate_up_proj_weight_quantizers.31 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.gate_up_proj_weight_quantizers.32 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.62e-01 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.gate_up_proj_weight_quantizers.33 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.13e-01 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.gate_up_proj_weight_quantizers.34 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.gate_up_proj_weight_quantizers.35 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.gate_up_proj_weight_quantizers.36 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.gate_up_proj_weight_quantizers.37 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.gate_up_proj_weight_quantizers.38 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.gate_up_proj_weight_quantizers.39 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.20e-01 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.gate_up_proj_weight_quantizers.40 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.25e-01 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.gate_up_proj_weight_quantizers.41 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.gate_up_proj_weight_quantizers.42 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.27e-01 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.gate_up_proj_weight_quantizers.43 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.gate_up_proj_weight_quantizers.44 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.gate_up_proj_weight_quantizers.45 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.gate_up_proj_weight_quantizers.46 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.gate_up_proj_weight_quantizers.47 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.42e-01 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.gate_up_proj_weight_quantizers.48 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.gate_up_proj_weight_quantizers.49 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.gate_up_proj_weight_quantizers.50 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.gate_up_proj_weight_quantizers.51 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.gate_up_proj_weight_quantizers.52 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.gate_up_proj_weight_quantizers.53 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.gate_up_proj_weight_quantizers.54 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.gate_up_proj_weight_quantizers.55 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.gate_up_proj_weight_quantizers.56 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.gate_up_proj_weight_quantizers.57 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.gate_up_proj_weight_quantizers.58 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.gate_up_proj_weight_quantizers.59 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.gate_up_proj_weight_quantizers.60 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.gate_up_proj_weight_quantizers.61 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.gate_up_proj_weight_quantizers.62 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.gate_up_proj_weight_quantizers.63 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.gate_up_proj_weight_quantizers.64 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.16e-01 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.gate_up_proj_weight_quantizers.65 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.gate_up_proj_weight_quantizers.66 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.gate_up_proj_weight_quantizers.67 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.77e-01 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.gate_up_proj_weight_quantizers.68 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.gate_up_proj_weight_quantizers.69 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.gate_up_proj_weight_quantizers.70 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.gate_up_proj_weight_quantizers.71 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.gate_up_proj_weight_quantizers.72 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.gate_up_proj_weight_quantizers.73 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.gate_up_proj_weight_quantizers.74 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.32e-01 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.gate_up_proj_weight_quantizers.75 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.99e-01 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.gate_up_proj_weight_quantizers.76 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.gate_up_proj_weight_quantizers.77 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.gate_up_proj_weight_quantizers.78 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.gate_up_proj_weight_quantizers.79 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.gate_up_proj_weight_quantizers.80 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.gate_up_proj_weight_quantizers.81 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.gate_up_proj_weight_quantizers.82 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.gate_up_proj_weight_quantizers.83 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.gate_up_proj_weight_quantizers.84 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.gate_up_proj_weight_quantizers.85 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.gate_up_proj_weight_quantizers.86 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.gate_up_proj_weight_quantizers.87 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.gate_up_proj_weight_quantizers.88 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.gate_up_proj_weight_quantizers.89 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.gate_up_proj_weight_quantizers.90 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.gate_up_proj_weight_quantizers.91 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.gate_up_proj_weight_quantizers.92 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.gate_up_proj_weight_quantizers.93 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.gate_up_proj_weight_quantizers.94 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.gate_up_proj_weight_quantizers.95 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.gate_up_proj_weight_quantizers.96 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.gate_up_proj_weight_quantizers.97 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.gate_up_proj_weight_quantizers.98 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.gate_up_proj_weight_quantizers.99 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.gate_up_proj_weight_quantizers.100 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.gate_up_proj_weight_quantizers.101 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.gate_up_proj_weight_quantizers.102 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.gate_up_proj_weight_quantizers.103 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.gate_up_proj_weight_quantizers.104 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.gate_up_proj_weight_quantizers.105 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.gate_up_proj_weight_quantizers.106 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.gate_up_proj_weight_quantizers.107 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.14e-01 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.gate_up_proj_weight_quantizers.108 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.gate_up_proj_weight_quantizers.109 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.gate_up_proj_weight_quantizers.110 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.gate_up_proj_weight_quantizers.111 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.gate_up_proj_weight_quantizers.112 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=4.65e-01 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.gate_up_proj_weight_quantizers.113 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.gate_up_proj_weight_quantizers.114 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.gate_up_proj_weight_quantizers.115 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.gate_up_proj_weight_quantizers.116 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.gate_up_proj_weight_quantizers.117 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.gate_up_proj_weight_quantizers.118 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.70e-01 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.gate_up_proj_weight_quantizers.119 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.53e-01 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.gate_up_proj_weight_quantizers.120 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.gate_up_proj_weight_quantizers.121 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.gate_up_proj_weight_quantizers.122 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.gate_up_proj_weight_quantizers.123 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.81e-01 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.gate_up_proj_weight_quantizers.124 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.gate_up_proj_weight_quantizers.125 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.14e-01 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.gate_up_proj_weight_quantizers.126 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.28e-01 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.gate_up_proj_weight_quantizers.127 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.gate_up_proj_weight_quantizers.128 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.gate_up_proj_weight_quantizers.129 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.gate_up_proj_weight_quantizers.130 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.gate_up_proj_weight_quantizers.131 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.gate_up_proj_weight_quantizers.132 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.gate_up_proj_weight_quantizers.133 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.gate_up_proj_weight_quantizers.134 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.gate_up_proj_weight_quantizers.135 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.16e-01 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.gate_up_proj_weight_quantizers.136 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.gate_up_proj_weight_quantizers.137 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.gate_up_proj_weight_quantizers.138 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.gate_up_proj_weight_quantizers.139 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.22e-01 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.gate_up_proj_weight_quantizers.140 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.gate_up_proj_weight_quantizers.141 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.gate_up_proj_weight_quantizers.142 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.gate_up_proj_weight_quantizers.143 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.gate_up_proj_weight_quantizers.144 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.gate_up_proj_weight_quantizers.145 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.gate_up_proj_weight_quantizers.146 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.15e-01 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.gate_up_proj_weight_quantizers.147 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.63e-01 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.gate_up_proj_weight_quantizers.148 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.gate_up_proj_weight_quantizers.149 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.gate_up_proj_weight_quantizers.150 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.gate_up_proj_weight_quantizers.151 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.gate_up_proj_weight_quantizers.152 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.42e-01 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.gate_up_proj_weight_quantizers.153 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.gate_up_proj_weight_quantizers.154 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.40e-01 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.gate_up_proj_weight_quantizers.155 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.gate_up_proj_weight_quantizers.156 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.gate_up_proj_weight_quantizers.157 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.05e-01 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.gate_up_proj_weight_quantizers.158 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.33e-01 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.gate_up_proj_weight_quantizers.159 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.gate_up_proj_weight_quantizers.160 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.16e-01 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.gate_up_proj_weight_quantizers.161 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.59e-01 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.gate_up_proj_weight_quantizers.162 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.gate_up_proj_weight_quantizers.163 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.gate_up_proj_weight_quantizers.164 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.gate_up_proj_weight_quantizers.165 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.gate_up_proj_weight_quantizers.166 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.16e-01 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.gate_up_proj_weight_quantizers.167 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.gate_up_proj_weight_quantizers.168 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.gate_up_proj_weight_quantizers.169 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.gate_up_proj_weight_quantizers.170 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.gate_up_proj_weight_quantizers.171 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.62e-01 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.gate_up_proj_weight_quantizers.172 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.gate_up_proj_weight_quantizers.173 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.gate_up_proj_weight_quantizers.174 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.gate_up_proj_weight_quantizers.175 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.gate_up_proj_weight_quantizers.176 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.gate_up_proj_weight_quantizers.177 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.gate_up_proj_weight_quantizers.178 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.13e-01 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.gate_up_proj_weight_quantizers.179 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.gate_up_proj_weight_quantizers.180 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.gate_up_proj_weight_quantizers.181 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.gate_up_proj_weight_quantizers.182 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.gate_up_proj_weight_quantizers.183 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.gate_up_proj_weight_quantizers.184 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.gate_up_proj_weight_quantizers.185 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.gate_up_proj_weight_quantizers.186 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.gate_up_proj_weight_quantizers.187 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.gate_up_proj_weight_quantizers.188 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.58e-01 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.gate_up_proj_weight_quantizers.189 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.gate_up_proj_weight_quantizers.190 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.gate_up_proj_weight_quantizers.191 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.65e-01 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.gate_up_proj_weight_quantizers.192 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.gate_up_proj_weight_quantizers.193 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.gate_up_proj_weight_quantizers.194 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.gate_up_proj_weight_quantizers.195 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.gate_up_proj_weight_quantizers.196 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.gate_up_proj_weight_quantizers.197 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.gate_up_proj_weight_quantizers.198 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.gate_up_proj_weight_quantizers.199 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.gate_up_proj_weight_quantizers.200 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.17e-01 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.gate_up_proj_weight_quantizers.201 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.70e-01 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.gate_up_proj_weight_quantizers.202 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.19e-01 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.gate_up_proj_weight_quantizers.203 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.60e-01 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.gate_up_proj_weight_quantizers.204 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.gate_up_proj_weight_quantizers.205 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.13e-01 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.gate_up_proj_weight_quantizers.206 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.17e-01 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.gate_up_proj_weight_quantizers.207 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.gate_up_proj_weight_quantizers.208 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.gate_up_proj_weight_quantizers.209 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.gate_up_proj_weight_quantizers.210 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.gate_up_proj_weight_quantizers.211 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.gate_up_proj_weight_quantizers.212 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.81e-01 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.gate_up_proj_weight_quantizers.213 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.gate_up_proj_weight_quantizers.214 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.gate_up_proj_weight_quantizers.215 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.41e-01 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.gate_up_proj_weight_quantizers.216 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.gate_up_proj_weight_quantizers.217 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.gate_up_proj_weight_quantizers.218 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.18e-01 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.gate_up_proj_weight_quantizers.219 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.gate_up_proj_weight_quantizers.220 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.gate_up_proj_weight_quantizers.221 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.gate_up_proj_weight_quantizers.222 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.gate_up_proj_weight_quantizers.223 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.22e-01 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.gate_up_proj_weight_quantizers.224 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.16e-01 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.gate_up_proj_weight_quantizers.225 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.gate_up_proj_weight_quantizers.226 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.91e-01 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.gate_up_proj_weight_quantizers.227 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.gate_up_proj_weight_quantizers.228 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.gate_up_proj_weight_quantizers.229 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.gate_up_proj_weight_quantizers.230 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.gate_up_proj_weight_quantizers.231 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.gate_up_proj_weight_quantizers.232 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.gate_up_proj_weight_quantizers.233 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.gate_up_proj_weight_quantizers.234 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.55e-01 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.gate_up_proj_weight_quantizers.235 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.53e-01 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.gate_up_proj_weight_quantizers.236 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.gate_up_proj_weight_quantizers.237 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.gate_up_proj_weight_quantizers.238 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.gate_up_proj_weight_quantizers.239 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.gate_up_proj_weight_quantizers.240 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.21e-01 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.gate_up_proj_weight_quantizers.241 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.14e-01 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.gate_up_proj_weight_quantizers.242 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.gate_up_proj_weight_quantizers.243 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.14e-01 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.gate_up_proj_weight_quantizers.244 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.gate_up_proj_weight_quantizers.245 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.14e-01 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.gate_up_proj_weight_quantizers.246 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.gate_up_proj_weight_quantizers.247 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.gate_up_proj_weight_quantizers.248 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.50e-01 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.gate_up_proj_weight_quantizers.249 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.36e-01 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.gate_up_proj_weight_quantizers.250 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.gate_up_proj_weight_quantizers.251 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.gate_up_proj_weight_quantizers.252 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.gate_up_proj_weight_quantizers.253 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.gate_up_proj_weight_quantizers.254 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.17e-01 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.gate_up_proj_weight_quantizers.255 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.down_proj_input_quantizer TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.24e+02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.down_proj_weight_quantizers.0 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.down_proj_weight_quantizers.1 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.down_proj_weight_quantizers.2 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.down_proj_weight_quantizers.3 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.down_proj_weight_quantizers.4 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.down_proj_weight_quantizers.5 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.down_proj_weight_quantizers.6 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.down_proj_weight_quantizers.7 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.down_proj_weight_quantizers.8 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.down_proj_weight_quantizers.9 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.down_proj_weight_quantizers.10 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.down_proj_weight_quantizers.11 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.45e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.down_proj_weight_quantizers.12 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.down_proj_weight_quantizers.13 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.down_proj_weight_quantizers.14 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.down_proj_weight_quantizers.15 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.down_proj_weight_quantizers.16 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.down_proj_weight_quantizers.17 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.down_proj_weight_quantizers.18 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.down_proj_weight_quantizers.19 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.down_proj_weight_quantizers.20 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.down_proj_weight_quantizers.21 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.down_proj_weight_quantizers.22 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.down_proj_weight_quantizers.23 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.down_proj_weight_quantizers.24 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.64e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.down_proj_weight_quantizers.25 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.down_proj_weight_quantizers.26 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.down_proj_weight_quantizers.27 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.59e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.down_proj_weight_quantizers.28 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.down_proj_weight_quantizers.29 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.down_proj_weight_quantizers.30 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.59e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.down_proj_weight_quantizers.31 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.down_proj_weight_quantizers.32 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.87e-01 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.down_proj_weight_quantizers.33 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.down_proj_weight_quantizers.34 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.down_proj_weight_quantizers.35 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.down_proj_weight_quantizers.36 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.down_proj_weight_quantizers.37 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.64e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.down_proj_weight_quantizers.38 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.down_proj_weight_quantizers.39 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.64e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.down_proj_weight_quantizers.40 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.down_proj_weight_quantizers.41 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.down_proj_weight_quantizers.42 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.13e-01 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.down_proj_weight_quantizers.43 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.down_proj_weight_quantizers.44 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.down_proj_weight_quantizers.45 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.down_proj_weight_quantizers.46 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.down_proj_weight_quantizers.47 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.74e-01 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.down_proj_weight_quantizers.48 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.down_proj_weight_quantizers.49 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.down_proj_weight_quantizers.50 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.down_proj_weight_quantizers.51 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.down_proj_weight_quantizers.52 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.down_proj_weight_quantizers.53 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.down_proj_weight_quantizers.54 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.down_proj_weight_quantizers.55 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.down_proj_weight_quantizers.56 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.down_proj_weight_quantizers.57 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.down_proj_weight_quantizers.58 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.down_proj_weight_quantizers.59 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.down_proj_weight_quantizers.60 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.down_proj_weight_quantizers.61 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.down_proj_weight_quantizers.62 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.down_proj_weight_quantizers.63 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.down_proj_weight_quantizers.64 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.down_proj_weight_quantizers.65 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.down_proj_weight_quantizers.66 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.down_proj_weight_quantizers.67 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.down_proj_weight_quantizers.68 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.down_proj_weight_quantizers.69 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.down_proj_weight_quantizers.70 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.down_proj_weight_quantizers.71 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.down_proj_weight_quantizers.72 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.14e-01 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.down_proj_weight_quantizers.73 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.down_proj_weight_quantizers.74 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.down_proj_weight_quantizers.75 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.down_proj_weight_quantizers.76 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.down_proj_weight_quantizers.77 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.down_proj_weight_quantizers.78 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.59e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.down_proj_weight_quantizers.79 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.down_proj_weight_quantizers.80 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.down_proj_weight_quantizers.81 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.down_proj_weight_quantizers.82 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.down_proj_weight_quantizers.83 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.down_proj_weight_quantizers.84 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.45e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.down_proj_weight_quantizers.85 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.down_proj_weight_quantizers.86 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.down_proj_weight_quantizers.87 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.down_proj_weight_quantizers.88 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.down_proj_weight_quantizers.89 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.down_proj_weight_quantizers.90 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.down_proj_weight_quantizers.91 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.down_proj_weight_quantizers.92 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.down_proj_weight_quantizers.93 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.down_proj_weight_quantizers.94 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.down_proj_weight_quantizers.95 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.down_proj_weight_quantizers.96 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.45e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.down_proj_weight_quantizers.97 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.down_proj_weight_quantizers.98 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.down_proj_weight_quantizers.99 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.down_proj_weight_quantizers.100 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.down_proj_weight_quantizers.101 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.down_proj_weight_quantizers.102 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.down_proj_weight_quantizers.103 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.down_proj_weight_quantizers.104 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.down_proj_weight_quantizers.105 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.38e-01 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.down_proj_weight_quantizers.106 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.down_proj_weight_quantizers.107 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.down_proj_weight_quantizers.108 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.down_proj_weight_quantizers.109 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.down_proj_weight_quantizers.110 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.down_proj_weight_quantizers.111 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.down_proj_weight_quantizers.112 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.32e-01 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.down_proj_weight_quantizers.113 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.down_proj_weight_quantizers.114 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.down_proj_weight_quantizers.115 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.down_proj_weight_quantizers.116 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.down_proj_weight_quantizers.117 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.down_proj_weight_quantizers.118 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.19e-01 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.down_proj_weight_quantizers.119 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.down_proj_weight_quantizers.120 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.45e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.down_proj_weight_quantizers.121 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.down_proj_weight_quantizers.122 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.down_proj_weight_quantizers.123 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.down_proj_weight_quantizers.124 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.59e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.down_proj_weight_quantizers.125 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.down_proj_weight_quantizers.126 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.down_proj_weight_quantizers.127 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.down_proj_weight_quantizers.128 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.54e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.down_proj_weight_quantizers.129 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.down_proj_weight_quantizers.130 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.down_proj_weight_quantizers.131 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.down_proj_weight_quantizers.132 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.down_proj_weight_quantizers.133 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.down_proj_weight_quantizers.134 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.down_proj_weight_quantizers.135 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.down_proj_weight_quantizers.136 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.54e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.down_proj_weight_quantizers.137 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.down_proj_weight_quantizers.138 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.down_proj_weight_quantizers.139 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.down_proj_weight_quantizers.140 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.down_proj_weight_quantizers.141 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.30e-01 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.down_proj_weight_quantizers.142 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.down_proj_weight_quantizers.143 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.down_proj_weight_quantizers.144 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.down_proj_weight_quantizers.145 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.down_proj_weight_quantizers.146 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.down_proj_weight_quantizers.147 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.down_proj_weight_quantizers.148 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.down_proj_weight_quantizers.149 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.down_proj_weight_quantizers.150 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.down_proj_weight_quantizers.151 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.27e-01 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.down_proj_weight_quantizers.152 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.down_proj_weight_quantizers.153 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.down_proj_weight_quantizers.154 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.down_proj_weight_quantizers.155 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.down_proj_weight_quantizers.156 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.down_proj_weight_quantizers.157 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.57e-01 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.down_proj_weight_quantizers.158 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.down_proj_weight_quantizers.159 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.down_proj_weight_quantizers.160 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.down_proj_weight_quantizers.161 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.31e-01 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.down_proj_weight_quantizers.162 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.down_proj_weight_quantizers.163 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.down_proj_weight_quantizers.164 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.down_proj_weight_quantizers.165 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.54e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.down_proj_weight_quantizers.166 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.down_proj_weight_quantizers.167 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.down_proj_weight_quantizers.168 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.down_proj_weight_quantizers.169 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.down_proj_weight_quantizers.170 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.down_proj_weight_quantizers.171 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.down_proj_weight_quantizers.172 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.down_proj_weight_quantizers.173 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.down_proj_weight_quantizers.174 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.down_proj_weight_quantizers.175 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.down_proj_weight_quantizers.176 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.down_proj_weight_quantizers.177 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.down_proj_weight_quantizers.178 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.54e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.down_proj_weight_quantizers.179 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.down_proj_weight_quantizers.180 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.down_proj_weight_quantizers.181 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.down_proj_weight_quantizers.182 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.down_proj_weight_quantizers.183 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.down_proj_weight_quantizers.184 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.down_proj_weight_quantizers.185 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.down_proj_weight_quantizers.186 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.down_proj_weight_quantizers.187 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.down_proj_weight_quantizers.188 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.down_proj_weight_quantizers.189 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.64e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.down_proj_weight_quantizers.190 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.50e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.down_proj_weight_quantizers.191 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.down_proj_weight_quantizers.192 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.down_proj_weight_quantizers.193 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.down_proj_weight_quantizers.194 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.down_proj_weight_quantizers.195 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.45e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.down_proj_weight_quantizers.196 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.down_proj_weight_quantizers.197 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.down_proj_weight_quantizers.198 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.down_proj_weight_quantizers.199 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.down_proj_weight_quantizers.200 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.down_proj_weight_quantizers.201 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.75e-01 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.down_proj_weight_quantizers.202 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.down_proj_weight_quantizers.203 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.down_proj_weight_quantizers.204 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.down_proj_weight_quantizers.205 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.down_proj_weight_quantizers.206 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.down_proj_weight_quantizers.207 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.down_proj_weight_quantizers.208 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.down_proj_weight_quantizers.209 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.down_proj_weight_quantizers.210 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.down_proj_weight_quantizers.211 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.down_proj_weight_quantizers.212 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.down_proj_weight_quantizers.213 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.down_proj_weight_quantizers.214 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.down_proj_weight_quantizers.215 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.down_proj_weight_quantizers.216 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.down_proj_weight_quantizers.217 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.down_proj_weight_quantizers.218 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.down_proj_weight_quantizers.219 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.54e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.down_proj_weight_quantizers.220 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.down_proj_weight_quantizers.221 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.down_proj_weight_quantizers.222 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.down_proj_weight_quantizers.223 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.down_proj_weight_quantizers.224 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.down_proj_weight_quantizers.225 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.down_proj_weight_quantizers.226 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.36e-01 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.down_proj_weight_quantizers.227 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.down_proj_weight_quantizers.228 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.down_proj_weight_quantizers.229 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.down_proj_weight_quantizers.230 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.down_proj_weight_quantizers.231 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.down_proj_weight_quantizers.232 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.down_proj_weight_quantizers.233 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.41e-01 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.down_proj_weight_quantizers.234 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.51e-01 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.down_proj_weight_quantizers.235 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.down_proj_weight_quantizers.236 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.down_proj_weight_quantizers.237 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.40e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.down_proj_weight_quantizers.238 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.down_proj_weight_quantizers.239 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.down_proj_weight_quantizers.240 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.down_proj_weight_quantizers.241 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.down_proj_weight_quantizers.242 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.down_proj_weight_quantizers.243 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.down_proj_weight_quantizers.244 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.down_proj_weight_quantizers.245 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.down_proj_weight_quantizers.246 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.down_proj_weight_quantizers.247 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.down_proj_weight_quantizers.248 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.59e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.down_proj_weight_quantizers.249 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.down_proj_weight_quantizers.250 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.down_proj_weight_quantizers.251 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.down_proj_weight_quantizers.252 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.down_proj_weight_quantizers.253 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.down_proj_weight_quantizers.254 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.down_proj_weight_quantizers.255 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.53.mlp.shared_experts.gate_proj.input_quantizer TensorQuantizer(disabled) +model.layers.53.mlp.shared_experts.gate_proj.output_quantizer TensorQuantizer(disabled) +model.layers.53.mlp.shared_experts.gate_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.53.mlp.shared_experts.up_proj.input_quantizer TensorQuantizer(disabled) +model.layers.53.mlp.shared_experts.up_proj.output_quantizer TensorQuantizer(disabled) +model.layers.53.mlp.shared_experts.up_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.53.mlp.shared_experts.down_proj.input_quantizer TensorQuantizer(disabled) +model.layers.53.mlp.shared_experts.down_proj.output_quantizer TensorQuantizer(disabled) +model.layers.53.mlp.shared_experts.down_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.54.self_attn.q_a_proj.input_quantizer TensorQuantizer(disabled) +model.layers.54.self_attn.q_a_proj.output_quantizer TensorQuantizer(disabled) +model.layers.54.self_attn.q_a_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.54.self_attn.q_b_proj.input_quantizer TensorQuantizer(disabled) +model.layers.54.self_attn.q_b_proj.output_quantizer TensorQuantizer(disabled) +model.layers.54.self_attn.q_b_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.54.self_attn.kv_a_proj_with_mqa.input_quantizer TensorQuantizer(disabled) +model.layers.54.self_attn.kv_a_proj_with_mqa.output_quantizer TensorQuantizer(disabled) +model.layers.54.self_attn.kv_a_proj_with_mqa.weight_quantizer TensorQuantizer(disabled) +model.layers.54.self_attn.kv_b_proj.input_quantizer TensorQuantizer(disabled) +model.layers.54.self_attn.kv_b_proj.output_quantizer TensorQuantizer(disabled) +model.layers.54.self_attn.kv_b_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.54.self_attn.o_proj.input_quantizer TensorQuantizer(disabled) +model.layers.54.self_attn.o_proj.output_quantizer TensorQuantizer(disabled) +model.layers.54.self_attn.o_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.54.self_attn.indexer.wq_b.input_quantizer TensorQuantizer(disabled) +model.layers.54.self_attn.indexer.wq_b.output_quantizer TensorQuantizer(disabled) +model.layers.54.self_attn.indexer.wq_b.weight_quantizer TensorQuantizer(disabled) +model.layers.54.self_attn.indexer.wk.input_quantizer TensorQuantizer(disabled) +model.layers.54.self_attn.indexer.wk.output_quantizer TensorQuantizer(disabled) +model.layers.54.self_attn.indexer.wk.weight_quantizer TensorQuantizer(disabled) +model.layers.54.self_attn.indexer.k_norm.input_quantizer TensorQuantizer(disabled) +model.layers.54.self_attn.indexer.k_norm.output_quantizer TensorQuantizer(disabled) +model.layers.54.self_attn.indexer.weights_proj.input_quantizer TensorQuantizer(disabled) +model.layers.54.self_attn.indexer.weights_proj.output_quantizer TensorQuantizer(disabled) +model.layers.54.self_attn.indexer.weights_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.54.self_attn.q_bmm_quantizer TensorQuantizer(disabled) +model.layers.54.self_attn.k_bmm_quantizer TensorQuantizer((4, 3) bit fake per-tensor amax=4.48e+02(const) calibrator=MaxCalibrator quant) +model.layers.54.self_attn.v_bmm_quantizer TensorQuantizer((4, 3) bit fake per-tensor amax=4.48e+02(const) calibrator=MaxCalibrator quant) +model.layers.54.self_attn.p_bmm_quantizer TensorQuantizer(disabled) +model.layers.54.mlp.experts.gate_up_proj_input_quantizer TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.38e+00 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.gate_up_proj_weight_quantizers.0 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.gate_up_proj_weight_quantizers.1 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.gate_up_proj_weight_quantizers.2 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.gate_up_proj_weight_quantizers.3 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.gate_up_proj_weight_quantizers.4 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.gate_up_proj_weight_quantizers.5 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.gate_up_proj_weight_quantizers.6 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.52e-01 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.gate_up_proj_weight_quantizers.7 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.gate_up_proj_weight_quantizers.8 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.17e-01 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.gate_up_proj_weight_quantizers.9 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.gate_up_proj_weight_quantizers.10 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.44e-01 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.gate_up_proj_weight_quantizers.11 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.gate_up_proj_weight_quantizers.12 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.gate_up_proj_weight_quantizers.13 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.40e-01 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.gate_up_proj_weight_quantizers.14 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.gate_up_proj_weight_quantizers.15 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.gate_up_proj_weight_quantizers.16 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.01e-01 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.gate_up_proj_weight_quantizers.17 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.16e-01 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.gate_up_proj_weight_quantizers.18 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.gate_up_proj_weight_quantizers.19 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.41e-01 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.gate_up_proj_weight_quantizers.20 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.gate_up_proj_weight_quantizers.21 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.15e-01 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.gate_up_proj_weight_quantizers.22 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.gate_up_proj_weight_quantizers.23 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.gate_up_proj_weight_quantizers.24 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.gate_up_proj_weight_quantizers.25 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.gate_up_proj_weight_quantizers.26 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.gate_up_proj_weight_quantizers.27 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.gate_up_proj_weight_quantizers.28 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.gate_up_proj_weight_quantizers.29 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.40e-01 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.gate_up_proj_weight_quantizers.30 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.83e-01 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.gate_up_proj_weight_quantizers.31 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.gate_up_proj_weight_quantizers.32 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.gate_up_proj_weight_quantizers.33 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.gate_up_proj_weight_quantizers.34 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.34e-01 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.gate_up_proj_weight_quantizers.35 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.16e-01 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.gate_up_proj_weight_quantizers.36 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.gate_up_proj_weight_quantizers.37 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.13e-01 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.gate_up_proj_weight_quantizers.38 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.gate_up_proj_weight_quantizers.39 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.66e-01 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.gate_up_proj_weight_quantizers.40 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.gate_up_proj_weight_quantizers.41 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.gate_up_proj_weight_quantizers.42 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.gate_up_proj_weight_quantizers.43 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.gate_up_proj_weight_quantizers.44 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.gate_up_proj_weight_quantizers.45 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.gate_up_proj_weight_quantizers.46 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.gate_up_proj_weight_quantizers.47 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.27e-01 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.gate_up_proj_weight_quantizers.48 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.64e-01 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.gate_up_proj_weight_quantizers.49 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.gate_up_proj_weight_quantizers.50 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.gate_up_proj_weight_quantizers.51 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.gate_up_proj_weight_quantizers.52 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.gate_up_proj_weight_quantizers.53 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.gate_up_proj_weight_quantizers.54 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.gate_up_proj_weight_quantizers.55 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.26e-01 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.gate_up_proj_weight_quantizers.56 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.25e-01 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.gate_up_proj_weight_quantizers.57 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.gate_up_proj_weight_quantizers.58 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.gate_up_proj_weight_quantizers.59 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.gate_up_proj_weight_quantizers.60 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.gate_up_proj_weight_quantizers.61 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.gate_up_proj_weight_quantizers.62 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.gate_up_proj_weight_quantizers.63 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.13e-01 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.gate_up_proj_weight_quantizers.64 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.gate_up_proj_weight_quantizers.65 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.gate_up_proj_weight_quantizers.66 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.63e-01 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.gate_up_proj_weight_quantizers.67 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.gate_up_proj_weight_quantizers.68 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.gate_up_proj_weight_quantizers.69 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.gate_up_proj_weight_quantizers.70 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.gate_up_proj_weight_quantizers.71 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.01e-01 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.gate_up_proj_weight_quantizers.72 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.gate_up_proj_weight_quantizers.73 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.gate_up_proj_weight_quantizers.74 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.gate_up_proj_weight_quantizers.75 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.gate_up_proj_weight_quantizers.76 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.62e-01 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.gate_up_proj_weight_quantizers.77 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.72e-01 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.gate_up_proj_weight_quantizers.78 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.29e-01 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.gate_up_proj_weight_quantizers.79 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.gate_up_proj_weight_quantizers.80 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.14e-01 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.gate_up_proj_weight_quantizers.81 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.gate_up_proj_weight_quantizers.82 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.gate_up_proj_weight_quantizers.83 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.37e-01 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.gate_up_proj_weight_quantizers.84 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.gate_up_proj_weight_quantizers.85 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.gate_up_proj_weight_quantizers.86 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.gate_up_proj_weight_quantizers.87 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.gate_up_proj_weight_quantizers.88 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.gate_up_proj_weight_quantizers.89 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.15e-01 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.gate_up_proj_weight_quantizers.90 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.gate_up_proj_weight_quantizers.91 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.gate_up_proj_weight_quantizers.92 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.22e-01 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.gate_up_proj_weight_quantizers.93 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.gate_up_proj_weight_quantizers.94 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.gate_up_proj_weight_quantizers.95 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.gate_up_proj_weight_quantizers.96 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.gate_up_proj_weight_quantizers.97 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.gate_up_proj_weight_quantizers.98 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.34e-01 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.gate_up_proj_weight_quantizers.99 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.gate_up_proj_weight_quantizers.100 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.gate_up_proj_weight_quantizers.101 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.gate_up_proj_weight_quantizers.102 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.gate_up_proj_weight_quantizers.103 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.gate_up_proj_weight_quantizers.104 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.gate_up_proj_weight_quantizers.105 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.20e-01 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.gate_up_proj_weight_quantizers.106 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.gate_up_proj_weight_quantizers.107 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.gate_up_proj_weight_quantizers.108 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.gate_up_proj_weight_quantizers.109 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.gate_up_proj_weight_quantizers.110 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.gate_up_proj_weight_quantizers.111 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.gate_up_proj_weight_quantizers.112 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.gate_up_proj_weight_quantizers.113 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.14e-01 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.gate_up_proj_weight_quantizers.114 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.gate_up_proj_weight_quantizers.115 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.gate_up_proj_weight_quantizers.116 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.gate_up_proj_weight_quantizers.117 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.28e-01 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.gate_up_proj_weight_quantizers.118 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.gate_up_proj_weight_quantizers.119 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.gate_up_proj_weight_quantizers.120 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.gate_up_proj_weight_quantizers.121 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.80e-01 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.gate_up_proj_weight_quantizers.122 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.13e-01 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.gate_up_proj_weight_quantizers.123 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.13e-01 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.gate_up_proj_weight_quantizers.124 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.gate_up_proj_weight_quantizers.125 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.gate_up_proj_weight_quantizers.126 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.gate_up_proj_weight_quantizers.127 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.16e-01 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.gate_up_proj_weight_quantizers.128 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.gate_up_proj_weight_quantizers.129 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.gate_up_proj_weight_quantizers.130 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.gate_up_proj_weight_quantizers.131 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.21e-01 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.gate_up_proj_weight_quantizers.132 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.gate_up_proj_weight_quantizers.133 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.gate_up_proj_weight_quantizers.134 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.gate_up_proj_weight_quantizers.135 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.gate_up_proj_weight_quantizers.136 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.gate_up_proj_weight_quantizers.137 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.gate_up_proj_weight_quantizers.138 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.gate_up_proj_weight_quantizers.139 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.gate_up_proj_weight_quantizers.140 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.gate_up_proj_weight_quantizers.141 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.gate_up_proj_weight_quantizers.142 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.gate_up_proj_weight_quantizers.143 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.gate_up_proj_weight_quantizers.144 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.39e-01 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.gate_up_proj_weight_quantizers.145 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.gate_up_proj_weight_quantizers.146 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.gate_up_proj_weight_quantizers.147 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.gate_up_proj_weight_quantizers.148 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.gate_up_proj_weight_quantizers.149 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.gate_up_proj_weight_quantizers.150 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.14e-01 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.gate_up_proj_weight_quantizers.151 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.20e-01 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.gate_up_proj_weight_quantizers.152 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.gate_up_proj_weight_quantizers.153 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.gate_up_proj_weight_quantizers.154 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.gate_up_proj_weight_quantizers.155 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.gate_up_proj_weight_quantizers.156 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.gate_up_proj_weight_quantizers.157 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.gate_up_proj_weight_quantizers.158 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.gate_up_proj_weight_quantizers.159 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.16e-01 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.gate_up_proj_weight_quantizers.160 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.gate_up_proj_weight_quantizers.161 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.92e-01 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.gate_up_proj_weight_quantizers.162 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.gate_up_proj_weight_quantizers.163 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.gate_up_proj_weight_quantizers.164 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.21e-01 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.gate_up_proj_weight_quantizers.165 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.gate_up_proj_weight_quantizers.166 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.gate_up_proj_weight_quantizers.167 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.12e-01 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.gate_up_proj_weight_quantizers.168 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.gate_up_proj_weight_quantizers.169 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.gate_up_proj_weight_quantizers.170 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.gate_up_proj_weight_quantizers.171 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.gate_up_proj_weight_quantizers.172 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.gate_up_proj_weight_quantizers.173 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.gate_up_proj_weight_quantizers.174 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.gate_up_proj_weight_quantizers.175 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.gate_up_proj_weight_quantizers.176 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.28e-01 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.gate_up_proj_weight_quantizers.177 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.gate_up_proj_weight_quantizers.178 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.17e-01 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.gate_up_proj_weight_quantizers.179 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.gate_up_proj_weight_quantizers.180 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.gate_up_proj_weight_quantizers.181 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.gate_up_proj_weight_quantizers.182 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.gate_up_proj_weight_quantizers.183 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.gate_up_proj_weight_quantizers.184 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.16e-01 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.gate_up_proj_weight_quantizers.185 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.gate_up_proj_weight_quantizers.186 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.30e-01 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.gate_up_proj_weight_quantizers.187 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.21e-01 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.gate_up_proj_weight_quantizers.188 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.gate_up_proj_weight_quantizers.189 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.gate_up_proj_weight_quantizers.190 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.gate_up_proj_weight_quantizers.191 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.gate_up_proj_weight_quantizers.192 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.15e-01 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.gate_up_proj_weight_quantizers.193 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.gate_up_proj_weight_quantizers.194 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.36e-01 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.gate_up_proj_weight_quantizers.195 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.gate_up_proj_weight_quantizers.196 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.gate_up_proj_weight_quantizers.197 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.gate_up_proj_weight_quantizers.198 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.gate_up_proj_weight_quantizers.199 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.gate_up_proj_weight_quantizers.200 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.gate_up_proj_weight_quantizers.201 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.gate_up_proj_weight_quantizers.202 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.81e-01 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.gate_up_proj_weight_quantizers.203 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.gate_up_proj_weight_quantizers.204 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.gate_up_proj_weight_quantizers.205 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.61e-01 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.gate_up_proj_weight_quantizers.206 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.gate_up_proj_weight_quantizers.207 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.gate_up_proj_weight_quantizers.208 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.26e-01 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.gate_up_proj_weight_quantizers.209 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.gate_up_proj_weight_quantizers.210 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.gate_up_proj_weight_quantizers.211 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.gate_up_proj_weight_quantizers.212 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.72e-01 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.gate_up_proj_weight_quantizers.213 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.19e-01 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.gate_up_proj_weight_quantizers.214 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.gate_up_proj_weight_quantizers.215 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.gate_up_proj_weight_quantizers.216 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.gate_up_proj_weight_quantizers.217 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.gate_up_proj_weight_quantizers.218 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.gate_up_proj_weight_quantizers.219 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.40e-01 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.gate_up_proj_weight_quantizers.220 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.gate_up_proj_weight_quantizers.221 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.45e-01 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.gate_up_proj_weight_quantizers.222 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.gate_up_proj_weight_quantizers.223 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.56e-01 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.gate_up_proj_weight_quantizers.224 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.gate_up_proj_weight_quantizers.225 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.gate_up_proj_weight_quantizers.226 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.30e-01 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.gate_up_proj_weight_quantizers.227 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.gate_up_proj_weight_quantizers.228 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.gate_up_proj_weight_quantizers.229 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.gate_up_proj_weight_quantizers.230 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.gate_up_proj_weight_quantizers.231 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.48e-01 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.gate_up_proj_weight_quantizers.232 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.gate_up_proj_weight_quantizers.233 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.gate_up_proj_weight_quantizers.234 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.gate_up_proj_weight_quantizers.235 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.gate_up_proj_weight_quantizers.236 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.gate_up_proj_weight_quantizers.237 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.18e-01 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.gate_up_proj_weight_quantizers.238 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.03e-01 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.gate_up_proj_weight_quantizers.239 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.gate_up_proj_weight_quantizers.240 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.gate_up_proj_weight_quantizers.241 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.31e-01 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.gate_up_proj_weight_quantizers.242 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.gate_up_proj_weight_quantizers.243 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.71e-01 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.gate_up_proj_weight_quantizers.244 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.gate_up_proj_weight_quantizers.245 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.gate_up_proj_weight_quantizers.246 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.gate_up_proj_weight_quantizers.247 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.19e-01 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.gate_up_proj_weight_quantizers.248 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.gate_up_proj_weight_quantizers.249 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.62e-01 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.gate_up_proj_weight_quantizers.250 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.gate_up_proj_weight_quantizers.251 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.26e-01 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.gate_up_proj_weight_quantizers.252 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.14e-01 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.gate_up_proj_weight_quantizers.253 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.gate_up_proj_weight_quantizers.254 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.gate_up_proj_weight_quantizers.255 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.01e-01 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.down_proj_input_quantizer TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.81e+01 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.down_proj_weight_quantizers.0 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.down_proj_weight_quantizers.1 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.down_proj_weight_quantizers.2 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.64e-02 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.down_proj_weight_quantizers.3 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.down_proj_weight_quantizers.4 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.down_proj_weight_quantizers.5 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.down_proj_weight_quantizers.6 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.down_proj_weight_quantizers.7 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.down_proj_weight_quantizers.8 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.down_proj_weight_quantizers.9 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.down_proj_weight_quantizers.10 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.03e-01 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.down_proj_weight_quantizers.11 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.down_proj_weight_quantizers.12 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.down_proj_weight_quantizers.13 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.down_proj_weight_quantizers.14 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.down_proj_weight_quantizers.15 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.down_proj_weight_quantizers.16 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.down_proj_weight_quantizers.17 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.down_proj_weight_quantizers.18 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.59e-02 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.down_proj_weight_quantizers.19 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.down_proj_weight_quantizers.20 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.down_proj_weight_quantizers.21 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.down_proj_weight_quantizers.22 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.down_proj_weight_quantizers.23 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.down_proj_weight_quantizers.24 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.down_proj_weight_quantizers.25 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.down_proj_weight_quantizers.26 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.down_proj_weight_quantizers.27 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.down_proj_weight_quantizers.28 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.down_proj_weight_quantizers.29 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.down_proj_weight_quantizers.30 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.down_proj_weight_quantizers.31 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.down_proj_weight_quantizers.32 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.down_proj_weight_quantizers.33 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.down_proj_weight_quantizers.34 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.down_proj_weight_quantizers.35 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.down_proj_weight_quantizers.36 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.down_proj_weight_quantizers.37 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.64e-02 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.down_proj_weight_quantizers.38 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.down_proj_weight_quantizers.39 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.down_proj_weight_quantizers.40 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.26e-01 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.down_proj_weight_quantizers.41 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.down_proj_weight_quantizers.42 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.down_proj_weight_quantizers.43 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.44e-01 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.down_proj_weight_quantizers.44 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.43e-01 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.down_proj_weight_quantizers.45 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.17e-01 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.down_proj_weight_quantizers.46 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.down_proj_weight_quantizers.47 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.down_proj_weight_quantizers.48 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.down_proj_weight_quantizers.49 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.down_proj_weight_quantizers.50 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.down_proj_weight_quantizers.51 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.down_proj_weight_quantizers.52 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.down_proj_weight_quantizers.53 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.down_proj_weight_quantizers.54 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.64e-02 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.down_proj_weight_quantizers.55 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.77e-01 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.down_proj_weight_quantizers.56 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.down_proj_weight_quantizers.57 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.18e-01 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.down_proj_weight_quantizers.58 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.down_proj_weight_quantizers.59 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.down_proj_weight_quantizers.60 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.down_proj_weight_quantizers.61 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.down_proj_weight_quantizers.62 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.48e-01 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.down_proj_weight_quantizers.63 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.41e-01 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.down_proj_weight_quantizers.64 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.down_proj_weight_quantizers.65 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.down_proj_weight_quantizers.66 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.down_proj_weight_quantizers.67 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.down_proj_weight_quantizers.68 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.down_proj_weight_quantizers.69 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.down_proj_weight_quantizers.70 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.down_proj_weight_quantizers.71 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.17e-01 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.down_proj_weight_quantizers.72 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.down_proj_weight_quantizers.73 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.down_proj_weight_quantizers.74 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.down_proj_weight_quantizers.75 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.down_proj_weight_quantizers.76 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.05e-01 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.down_proj_weight_quantizers.77 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.03e-01 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.down_proj_weight_quantizers.78 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.down_proj_weight_quantizers.79 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.down_proj_weight_quantizers.80 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.down_proj_weight_quantizers.81 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.down_proj_weight_quantizers.82 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.down_proj_weight_quantizers.83 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.44e-01 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.down_proj_weight_quantizers.84 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.down_proj_weight_quantizers.85 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.down_proj_weight_quantizers.86 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.down_proj_weight_quantizers.87 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.down_proj_weight_quantizers.88 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.down_proj_weight_quantizers.89 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.down_proj_weight_quantizers.90 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.down_proj_weight_quantizers.91 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.down_proj_weight_quantizers.92 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.down_proj_weight_quantizers.93 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.down_proj_weight_quantizers.94 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.down_proj_weight_quantizers.95 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.down_proj_weight_quantizers.96 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.down_proj_weight_quantizers.97 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.down_proj_weight_quantizers.98 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.35e-01 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.down_proj_weight_quantizers.99 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.down_proj_weight_quantizers.100 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.down_proj_weight_quantizers.101 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.down_proj_weight_quantizers.102 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.down_proj_weight_quantizers.103 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.down_proj_weight_quantizers.104 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.down_proj_weight_quantizers.105 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.down_proj_weight_quantizers.106 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.down_proj_weight_quantizers.107 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.down_proj_weight_quantizers.108 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.down_proj_weight_quantizers.109 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.down_proj_weight_quantizers.110 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.down_proj_weight_quantizers.111 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.down_proj_weight_quantizers.112 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.down_proj_weight_quantizers.113 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.down_proj_weight_quantizers.114 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.down_proj_weight_quantizers.115 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.down_proj_weight_quantizers.116 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.down_proj_weight_quantizers.117 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.down_proj_weight_quantizers.118 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.down_proj_weight_quantizers.119 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.down_proj_weight_quantizers.120 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.down_proj_weight_quantizers.121 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.54e-01 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.down_proj_weight_quantizers.122 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.down_proj_weight_quantizers.123 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.down_proj_weight_quantizers.124 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.down_proj_weight_quantizers.125 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.54e-01 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.down_proj_weight_quantizers.126 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.down_proj_weight_quantizers.127 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.down_proj_weight_quantizers.128 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.down_proj_weight_quantizers.129 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.35e-02 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.down_proj_weight_quantizers.130 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.down_proj_weight_quantizers.131 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.down_proj_weight_quantizers.132 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.down_proj_weight_quantizers.133 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.down_proj_weight_quantizers.134 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.down_proj_weight_quantizers.135 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.down_proj_weight_quantizers.136 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.down_proj_weight_quantizers.137 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.down_proj_weight_quantizers.138 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.64e-02 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.down_proj_weight_quantizers.139 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.18e-01 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.down_proj_weight_quantizers.140 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.down_proj_weight_quantizers.141 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.down_proj_weight_quantizers.142 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.down_proj_weight_quantizers.143 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.down_proj_weight_quantizers.144 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.down_proj_weight_quantizers.145 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.down_proj_weight_quantizers.146 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.down_proj_weight_quantizers.147 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.down_proj_weight_quantizers.148 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.27e-01 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.down_proj_weight_quantizers.149 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.50e-02 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.down_proj_weight_quantizers.150 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.57e-01 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.down_proj_weight_quantizers.151 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.down_proj_weight_quantizers.152 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.down_proj_weight_quantizers.153 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.down_proj_weight_quantizers.154 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.down_proj_weight_quantizers.155 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.down_proj_weight_quantizers.156 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.down_proj_weight_quantizers.157 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.down_proj_weight_quantizers.158 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.down_proj_weight_quantizers.159 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.down_proj_weight_quantizers.160 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.down_proj_weight_quantizers.161 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.88e-01 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.down_proj_weight_quantizers.162 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.down_proj_weight_quantizers.163 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.down_proj_weight_quantizers.164 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.down_proj_weight_quantizers.165 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.64e-02 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.down_proj_weight_quantizers.166 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.down_proj_weight_quantizers.167 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.down_proj_weight_quantizers.168 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.down_proj_weight_quantizers.169 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.down_proj_weight_quantizers.170 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.down_proj_weight_quantizers.171 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.down_proj_weight_quantizers.172 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.down_proj_weight_quantizers.173 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.down_proj_weight_quantizers.174 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.down_proj_weight_quantizers.175 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.down_proj_weight_quantizers.176 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.down_proj_weight_quantizers.177 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.down_proj_weight_quantizers.178 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.down_proj_weight_quantizers.179 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.down_proj_weight_quantizers.180 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.down_proj_weight_quantizers.181 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.64e-02 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.down_proj_weight_quantizers.182 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.down_proj_weight_quantizers.183 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.down_proj_weight_quantizers.184 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.down_proj_weight_quantizers.185 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.down_proj_weight_quantizers.186 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.down_proj_weight_quantizers.187 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.down_proj_weight_quantizers.188 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.down_proj_weight_quantizers.189 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.down_proj_weight_quantizers.190 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.down_proj_weight_quantizers.191 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.down_proj_weight_quantizers.192 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.24e-01 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.down_proj_weight_quantizers.193 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.down_proj_weight_quantizers.194 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.53e-01 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.down_proj_weight_quantizers.195 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.down_proj_weight_quantizers.196 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.down_proj_weight_quantizers.197 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.down_proj_weight_quantizers.198 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.down_proj_weight_quantizers.199 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.down_proj_weight_quantizers.200 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.down_proj_weight_quantizers.201 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.down_proj_weight_quantizers.202 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.down_proj_weight_quantizers.203 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.down_proj_weight_quantizers.204 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.down_proj_weight_quantizers.205 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.down_proj_weight_quantizers.206 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.64e-02 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.down_proj_weight_quantizers.207 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.down_proj_weight_quantizers.208 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.down_proj_weight_quantizers.209 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.54e-02 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.down_proj_weight_quantizers.210 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.down_proj_weight_quantizers.211 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.down_proj_weight_quantizers.212 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.01e-01 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.down_proj_weight_quantizers.213 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.down_proj_weight_quantizers.214 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.down_proj_weight_quantizers.215 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.down_proj_weight_quantizers.216 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.down_proj_weight_quantizers.217 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.down_proj_weight_quantizers.218 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.41e-01 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.down_proj_weight_quantizers.219 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.down_proj_weight_quantizers.220 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.down_proj_weight_quantizers.221 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.down_proj_weight_quantizers.222 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.down_proj_weight_quantizers.223 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.63e-01 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.down_proj_weight_quantizers.224 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.down_proj_weight_quantizers.225 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.down_proj_weight_quantizers.226 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.down_proj_weight_quantizers.227 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.down_proj_weight_quantizers.228 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.down_proj_weight_quantizers.229 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.down_proj_weight_quantizers.230 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.down_proj_weight_quantizers.231 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.down_proj_weight_quantizers.232 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.down_proj_weight_quantizers.233 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.down_proj_weight_quantizers.234 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.down_proj_weight_quantizers.235 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.down_proj_weight_quantizers.236 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.down_proj_weight_quantizers.237 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.down_proj_weight_quantizers.238 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.down_proj_weight_quantizers.239 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.down_proj_weight_quantizers.240 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.down_proj_weight_quantizers.241 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.down_proj_weight_quantizers.242 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.down_proj_weight_quantizers.243 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.41e-01 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.down_proj_weight_quantizers.244 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.down_proj_weight_quantizers.245 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.down_proj_weight_quantizers.246 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.down_proj_weight_quantizers.247 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.down_proj_weight_quantizers.248 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.down_proj_weight_quantizers.249 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.down_proj_weight_quantizers.250 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.down_proj_weight_quantizers.251 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.down_proj_weight_quantizers.252 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.79e-01 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.down_proj_weight_quantizers.253 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.down_proj_weight_quantizers.254 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.down_proj_weight_quantizers.255 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.54.mlp.shared_experts.gate_proj.input_quantizer TensorQuantizer(disabled) +model.layers.54.mlp.shared_experts.gate_proj.output_quantizer TensorQuantizer(disabled) +model.layers.54.mlp.shared_experts.gate_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.54.mlp.shared_experts.up_proj.input_quantizer TensorQuantizer(disabled) +model.layers.54.mlp.shared_experts.up_proj.output_quantizer TensorQuantizer(disabled) +model.layers.54.mlp.shared_experts.up_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.54.mlp.shared_experts.down_proj.input_quantizer TensorQuantizer(disabled) +model.layers.54.mlp.shared_experts.down_proj.output_quantizer TensorQuantizer(disabled) +model.layers.54.mlp.shared_experts.down_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.55.self_attn.q_a_proj.input_quantizer TensorQuantizer(disabled) +model.layers.55.self_attn.q_a_proj.output_quantizer TensorQuantizer(disabled) +model.layers.55.self_attn.q_a_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.55.self_attn.q_b_proj.input_quantizer TensorQuantizer(disabled) +model.layers.55.self_attn.q_b_proj.output_quantizer TensorQuantizer(disabled) +model.layers.55.self_attn.q_b_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.55.self_attn.kv_a_proj_with_mqa.input_quantizer TensorQuantizer(disabled) +model.layers.55.self_attn.kv_a_proj_with_mqa.output_quantizer TensorQuantizer(disabled) +model.layers.55.self_attn.kv_a_proj_with_mqa.weight_quantizer TensorQuantizer(disabled) +model.layers.55.self_attn.kv_b_proj.input_quantizer TensorQuantizer(disabled) +model.layers.55.self_attn.kv_b_proj.output_quantizer TensorQuantizer(disabled) +model.layers.55.self_attn.kv_b_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.55.self_attn.o_proj.input_quantizer TensorQuantizer(disabled) +model.layers.55.self_attn.o_proj.output_quantizer TensorQuantizer(disabled) +model.layers.55.self_attn.o_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.55.self_attn.q_bmm_quantizer TensorQuantizer(disabled) +model.layers.55.self_attn.k_bmm_quantizer TensorQuantizer((4, 3) bit fake per-tensor amax=4.48e+02(const) calibrator=MaxCalibrator quant) +model.layers.55.self_attn.v_bmm_quantizer TensorQuantizer((4, 3) bit fake per-tensor amax=4.48e+02(const) calibrator=MaxCalibrator quant) +model.layers.55.self_attn.p_bmm_quantizer TensorQuantizer(disabled) +model.layers.55.mlp.experts.gate_up_proj_input_quantizer TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.62e+00 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.gate_up_proj_weight_quantizers.0 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.73e-01 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.gate_up_proj_weight_quantizers.1 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.gate_up_proj_weight_quantizers.2 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.51e-01 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.gate_up_proj_weight_quantizers.3 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.gate_up_proj_weight_quantizers.4 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.gate_up_proj_weight_quantizers.5 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.gate_up_proj_weight_quantizers.6 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.gate_up_proj_weight_quantizers.7 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.gate_up_proj_weight_quantizers.8 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.88e-01 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.gate_up_proj_weight_quantizers.9 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.16e-01 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.gate_up_proj_weight_quantizers.10 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.25e-01 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.gate_up_proj_weight_quantizers.11 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.gate_up_proj_weight_quantizers.12 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.gate_up_proj_weight_quantizers.13 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.gate_up_proj_weight_quantizers.14 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.gate_up_proj_weight_quantizers.15 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.gate_up_proj_weight_quantizers.16 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.gate_up_proj_weight_quantizers.17 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.gate_up_proj_weight_quantizers.18 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.53e-01 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.gate_up_proj_weight_quantizers.19 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.gate_up_proj_weight_quantizers.20 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.gate_up_proj_weight_quantizers.21 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.gate_up_proj_weight_quantizers.22 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.gate_up_proj_weight_quantizers.23 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.gate_up_proj_weight_quantizers.24 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.gate_up_proj_weight_quantizers.25 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.gate_up_proj_weight_quantizers.26 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.60e-01 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.gate_up_proj_weight_quantizers.27 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.gate_up_proj_weight_quantizers.28 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.gate_up_proj_weight_quantizers.29 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.gate_up_proj_weight_quantizers.30 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.23e-01 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.gate_up_proj_weight_quantizers.31 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.gate_up_proj_weight_quantizers.32 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.gate_up_proj_weight_quantizers.33 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.gate_up_proj_weight_quantizers.34 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.gate_up_proj_weight_quantizers.35 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.17e-01 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.gate_up_proj_weight_quantizers.36 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.gate_up_proj_weight_quantizers.37 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.gate_up_proj_weight_quantizers.38 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.gate_up_proj_weight_quantizers.39 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.gate_up_proj_weight_quantizers.40 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.20e-01 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.gate_up_proj_weight_quantizers.41 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.gate_up_proj_weight_quantizers.42 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.gate_up_proj_weight_quantizers.43 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.gate_up_proj_weight_quantizers.44 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.57e-01 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.gate_up_proj_weight_quantizers.45 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.gate_up_proj_weight_quantizers.46 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.gate_up_proj_weight_quantizers.47 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.gate_up_proj_weight_quantizers.48 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.gate_up_proj_weight_quantizers.49 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.gate_up_proj_weight_quantizers.50 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.gate_up_proj_weight_quantizers.51 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.gate_up_proj_weight_quantizers.52 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.13e-01 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.gate_up_proj_weight_quantizers.53 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.gate_up_proj_weight_quantizers.54 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.gate_up_proj_weight_quantizers.55 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.gate_up_proj_weight_quantizers.56 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.38e-01 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.gate_up_proj_weight_quantizers.57 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.gate_up_proj_weight_quantizers.58 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.gate_up_proj_weight_quantizers.59 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.13e-01 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.gate_up_proj_weight_quantizers.60 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.17e-01 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.gate_up_proj_weight_quantizers.61 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.gate_up_proj_weight_quantizers.62 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.gate_up_proj_weight_quantizers.63 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.gate_up_proj_weight_quantizers.64 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.47e-01 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.gate_up_proj_weight_quantizers.65 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.gate_up_proj_weight_quantizers.66 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.gate_up_proj_weight_quantizers.67 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.gate_up_proj_weight_quantizers.68 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.gate_up_proj_weight_quantizers.69 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.gate_up_proj_weight_quantizers.70 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.gate_up_proj_weight_quantizers.71 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.58e-01 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.gate_up_proj_weight_quantizers.72 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.19e-01 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.gate_up_proj_weight_quantizers.73 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.gate_up_proj_weight_quantizers.74 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.gate_up_proj_weight_quantizers.75 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.gate_up_proj_weight_quantizers.76 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.14e-01 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.gate_up_proj_weight_quantizers.77 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.gate_up_proj_weight_quantizers.78 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.gate_up_proj_weight_quantizers.79 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.gate_up_proj_weight_quantizers.80 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.gate_up_proj_weight_quantizers.81 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.99e-01 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.gate_up_proj_weight_quantizers.82 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.gate_up_proj_weight_quantizers.83 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.gate_up_proj_weight_quantizers.84 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.gate_up_proj_weight_quantizers.85 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.gate_up_proj_weight_quantizers.86 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.20e-01 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.gate_up_proj_weight_quantizers.87 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.gate_up_proj_weight_quantizers.88 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.gate_up_proj_weight_quantizers.89 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.gate_up_proj_weight_quantizers.90 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.32e-01 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.gate_up_proj_weight_quantizers.91 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.gate_up_proj_weight_quantizers.92 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.gate_up_proj_weight_quantizers.93 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.gate_up_proj_weight_quantizers.94 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.gate_up_proj_weight_quantizers.95 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.gate_up_proj_weight_quantizers.96 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.gate_up_proj_weight_quantizers.97 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.gate_up_proj_weight_quantizers.98 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.gate_up_proj_weight_quantizers.99 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.gate_up_proj_weight_quantizers.100 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.gate_up_proj_weight_quantizers.101 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.gate_up_proj_weight_quantizers.102 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.gate_up_proj_weight_quantizers.103 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.gate_up_proj_weight_quantizers.104 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.gate_up_proj_weight_quantizers.105 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.gate_up_proj_weight_quantizers.106 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.gate_up_proj_weight_quantizers.107 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.gate_up_proj_weight_quantizers.108 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.gate_up_proj_weight_quantizers.109 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.84e-01 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.gate_up_proj_weight_quantizers.110 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.gate_up_proj_weight_quantizers.111 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.gate_up_proj_weight_quantizers.112 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.gate_up_proj_weight_quantizers.113 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.gate_up_proj_weight_quantizers.114 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.gate_up_proj_weight_quantizers.115 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.gate_up_proj_weight_quantizers.116 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.28e-01 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.gate_up_proj_weight_quantizers.117 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.gate_up_proj_weight_quantizers.118 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.gate_up_proj_weight_quantizers.119 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.gate_up_proj_weight_quantizers.120 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.gate_up_proj_weight_quantizers.121 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.20e-01 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.gate_up_proj_weight_quantizers.122 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.gate_up_proj_weight_quantizers.123 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.gate_up_proj_weight_quantizers.124 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.13e-01 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.gate_up_proj_weight_quantizers.125 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.gate_up_proj_weight_quantizers.126 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.gate_up_proj_weight_quantizers.127 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.gate_up_proj_weight_quantizers.128 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.gate_up_proj_weight_quantizers.129 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.gate_up_proj_weight_quantizers.130 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.gate_up_proj_weight_quantizers.131 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.gate_up_proj_weight_quantizers.132 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.17e-01 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.gate_up_proj_weight_quantizers.133 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.gate_up_proj_weight_quantizers.134 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.gate_up_proj_weight_quantizers.135 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.gate_up_proj_weight_quantizers.136 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.gate_up_proj_weight_quantizers.137 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.gate_up_proj_weight_quantizers.138 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.gate_up_proj_weight_quantizers.139 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.gate_up_proj_weight_quantizers.140 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.gate_up_proj_weight_quantizers.141 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.53e-01 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.gate_up_proj_weight_quantizers.142 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.gate_up_proj_weight_quantizers.143 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.gate_up_proj_weight_quantizers.144 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.gate_up_proj_weight_quantizers.145 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.gate_up_proj_weight_quantizers.146 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.23e-01 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.gate_up_proj_weight_quantizers.147 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.gate_up_proj_weight_quantizers.148 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.gate_up_proj_weight_quantizers.149 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=6.84e-01 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.gate_up_proj_weight_quantizers.150 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.gate_up_proj_weight_quantizers.151 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.gate_up_proj_weight_quantizers.152 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.20e-01 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.gate_up_proj_weight_quantizers.153 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.gate_up_proj_weight_quantizers.154 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.gate_up_proj_weight_quantizers.155 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.gate_up_proj_weight_quantizers.156 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.gate_up_proj_weight_quantizers.157 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.gate_up_proj_weight_quantizers.158 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.gate_up_proj_weight_quantizers.159 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.gate_up_proj_weight_quantizers.160 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.gate_up_proj_weight_quantizers.161 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.gate_up_proj_weight_quantizers.162 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.gate_up_proj_weight_quantizers.163 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.13e-01 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.gate_up_proj_weight_quantizers.164 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.35e-01 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.gate_up_proj_weight_quantizers.165 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.gate_up_proj_weight_quantizers.166 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.gate_up_proj_weight_quantizers.167 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.gate_up_proj_weight_quantizers.168 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.gate_up_proj_weight_quantizers.169 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.gate_up_proj_weight_quantizers.170 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.31e-01 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.gate_up_proj_weight_quantizers.171 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.gate_up_proj_weight_quantizers.172 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.54e-01 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.gate_up_proj_weight_quantizers.173 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.gate_up_proj_weight_quantizers.174 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.13e-01 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.gate_up_proj_weight_quantizers.175 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.gate_up_proj_weight_quantizers.176 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.gate_up_proj_weight_quantizers.177 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.gate_up_proj_weight_quantizers.178 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.gate_up_proj_weight_quantizers.179 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.gate_up_proj_weight_quantizers.180 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.gate_up_proj_weight_quantizers.181 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.gate_up_proj_weight_quantizers.182 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.gate_up_proj_weight_quantizers.183 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.gate_up_proj_weight_quantizers.184 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.gate_up_proj_weight_quantizers.185 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.gate_up_proj_weight_quantizers.186 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.gate_up_proj_weight_quantizers.187 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.gate_up_proj_weight_quantizers.188 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.gate_up_proj_weight_quantizers.189 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.gate_up_proj_weight_quantizers.190 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.20e-01 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.gate_up_proj_weight_quantizers.191 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.gate_up_proj_weight_quantizers.192 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.gate_up_proj_weight_quantizers.193 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.gate_up_proj_weight_quantizers.194 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.gate_up_proj_weight_quantizers.195 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.gate_up_proj_weight_quantizers.196 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.gate_up_proj_weight_quantizers.197 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.gate_up_proj_weight_quantizers.198 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.gate_up_proj_weight_quantizers.199 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.28e-01 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.gate_up_proj_weight_quantizers.200 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.gate_up_proj_weight_quantizers.201 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.gate_up_proj_weight_quantizers.202 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.gate_up_proj_weight_quantizers.203 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.gate_up_proj_weight_quantizers.204 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.gate_up_proj_weight_quantizers.205 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.gate_up_proj_weight_quantizers.206 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.gate_up_proj_weight_quantizers.207 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.gate_up_proj_weight_quantizers.208 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.gate_up_proj_weight_quantizers.209 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.gate_up_proj_weight_quantizers.210 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.gate_up_proj_weight_quantizers.211 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.59e-02 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.gate_up_proj_weight_quantizers.212 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.gate_up_proj_weight_quantizers.213 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.gate_up_proj_weight_quantizers.214 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.gate_up_proj_weight_quantizers.215 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.gate_up_proj_weight_quantizers.216 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.gate_up_proj_weight_quantizers.217 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.gate_up_proj_weight_quantizers.218 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.gate_up_proj_weight_quantizers.219 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.gate_up_proj_weight_quantizers.220 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.gate_up_proj_weight_quantizers.221 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.gate_up_proj_weight_quantizers.222 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.gate_up_proj_weight_quantizers.223 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.gate_up_proj_weight_quantizers.224 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.gate_up_proj_weight_quantizers.225 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.gate_up_proj_weight_quantizers.226 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.gate_up_proj_weight_quantizers.227 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.gate_up_proj_weight_quantizers.228 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.19e-01 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.gate_up_proj_weight_quantizers.229 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.gate_up_proj_weight_quantizers.230 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.gate_up_proj_weight_quantizers.231 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.gate_up_proj_weight_quantizers.232 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.gate_up_proj_weight_quantizers.233 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.gate_up_proj_weight_quantizers.234 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.gate_up_proj_weight_quantizers.235 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.gate_up_proj_weight_quantizers.236 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.gate_up_proj_weight_quantizers.237 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.55e-01 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.gate_up_proj_weight_quantizers.238 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.gate_up_proj_weight_quantizers.239 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.gate_up_proj_weight_quantizers.240 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.26e-01 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.gate_up_proj_weight_quantizers.241 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.gate_up_proj_weight_quantizers.242 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.15e-01 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.gate_up_proj_weight_quantizers.243 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.42e-01 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.gate_up_proj_weight_quantizers.244 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.gate_up_proj_weight_quantizers.245 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.gate_up_proj_weight_quantizers.246 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.gate_up_proj_weight_quantizers.247 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.gate_up_proj_weight_quantizers.248 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.gate_up_proj_weight_quantizers.249 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.gate_up_proj_weight_quantizers.250 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.gate_up_proj_weight_quantizers.251 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.gate_up_proj_weight_quantizers.252 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.gate_up_proj_weight_quantizers.253 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.gate_up_proj_weight_quantizers.254 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.gate_up_proj_weight_quantizers.255 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.down_proj_input_quantizer TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.92e+01 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.down_proj_weight_quantizers.0 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.98e-01 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.down_proj_weight_quantizers.1 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.down_proj_weight_quantizers.2 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.45e-01 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.down_proj_weight_quantizers.3 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.down_proj_weight_quantizers.4 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.59e-02 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.down_proj_weight_quantizers.5 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.down_proj_weight_quantizers.6 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.down_proj_weight_quantizers.7 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.55e-01 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.down_proj_weight_quantizers.8 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.down_proj_weight_quantizers.9 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.down_proj_weight_quantizers.10 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.down_proj_weight_quantizers.11 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.down_proj_weight_quantizers.12 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.down_proj_weight_quantizers.13 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.35e-02 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.down_proj_weight_quantizers.14 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.down_proj_weight_quantizers.15 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.down_proj_weight_quantizers.16 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.down_proj_weight_quantizers.17 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.31e-01 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.down_proj_weight_quantizers.18 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.down_proj_weight_quantizers.19 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.down_proj_weight_quantizers.20 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.down_proj_weight_quantizers.21 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.down_proj_weight_quantizers.22 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.16e-01 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.down_proj_weight_quantizers.23 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.64e-02 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.down_proj_weight_quantizers.24 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.down_proj_weight_quantizers.25 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.13e-01 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.down_proj_weight_quantizers.26 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.55e-01 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.down_proj_weight_quantizers.27 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.down_proj_weight_quantizers.28 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.down_proj_weight_quantizers.29 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.down_proj_weight_quantizers.30 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.27e-01 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.down_proj_weight_quantizers.31 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.down_proj_weight_quantizers.32 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.73e-01 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.down_proj_weight_quantizers.33 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.59e-02 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.down_proj_weight_quantizers.34 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.01e-01 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.down_proj_weight_quantizers.35 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.down_proj_weight_quantizers.36 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.45e-02 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.down_proj_weight_quantizers.37 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.down_proj_weight_quantizers.38 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.down_proj_weight_quantizers.39 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.down_proj_weight_quantizers.40 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.down_proj_weight_quantizers.41 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.45e-02 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.down_proj_weight_quantizers.42 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.down_proj_weight_quantizers.43 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.59e-02 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.down_proj_weight_quantizers.44 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.down_proj_weight_quantizers.45 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.down_proj_weight_quantizers.46 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.down_proj_weight_quantizers.47 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.down_proj_weight_quantizers.48 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.down_proj_weight_quantizers.49 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.down_proj_weight_quantizers.50 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.down_proj_weight_quantizers.51 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.down_proj_weight_quantizers.52 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.down_proj_weight_quantizers.53 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.down_proj_weight_quantizers.54 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.54e-02 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.down_proj_weight_quantizers.55 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.down_proj_weight_quantizers.56 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.down_proj_weight_quantizers.57 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.down_proj_weight_quantizers.58 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.down_proj_weight_quantizers.59 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.down_proj_weight_quantizers.60 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.down_proj_weight_quantizers.61 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.down_proj_weight_quantizers.62 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.54e-02 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.down_proj_weight_quantizers.63 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.64e-02 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.down_proj_weight_quantizers.64 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.78e-01 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.down_proj_weight_quantizers.65 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.down_proj_weight_quantizers.66 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.down_proj_weight_quantizers.67 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.down_proj_weight_quantizers.68 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.64e-02 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.down_proj_weight_quantizers.69 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.down_proj_weight_quantizers.70 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.down_proj_weight_quantizers.71 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.down_proj_weight_quantizers.72 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.down_proj_weight_quantizers.73 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.down_proj_weight_quantizers.74 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.down_proj_weight_quantizers.75 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.down_proj_weight_quantizers.76 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.down_proj_weight_quantizers.77 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.down_proj_weight_quantizers.78 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.down_proj_weight_quantizers.79 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.down_proj_weight_quantizers.80 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.down_proj_weight_quantizers.81 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.down_proj_weight_quantizers.82 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.down_proj_weight_quantizers.83 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.down_proj_weight_quantizers.84 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.down_proj_weight_quantizers.85 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.66e-01 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.down_proj_weight_quantizers.86 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.down_proj_weight_quantizers.87 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.down_proj_weight_quantizers.88 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.down_proj_weight_quantizers.89 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.down_proj_weight_quantizers.90 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.78e-01 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.down_proj_weight_quantizers.91 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.down_proj_weight_quantizers.92 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.down_proj_weight_quantizers.93 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.down_proj_weight_quantizers.94 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.97e-01 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.down_proj_weight_quantizers.95 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.down_proj_weight_quantizers.96 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.down_proj_weight_quantizers.97 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.down_proj_weight_quantizers.98 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.down_proj_weight_quantizers.99 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.down_proj_weight_quantizers.100 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.down_proj_weight_quantizers.101 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.down_proj_weight_quantizers.102 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.down_proj_weight_quantizers.103 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.down_proj_weight_quantizers.104 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.down_proj_weight_quantizers.105 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.down_proj_weight_quantizers.106 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.down_proj_weight_quantizers.107 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.down_proj_weight_quantizers.108 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.down_proj_weight_quantizers.109 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.down_proj_weight_quantizers.110 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.down_proj_weight_quantizers.111 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.down_proj_weight_quantizers.112 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.down_proj_weight_quantizers.113 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.down_proj_weight_quantizers.114 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.down_proj_weight_quantizers.115 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.down_proj_weight_quantizers.116 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.61e-01 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.down_proj_weight_quantizers.117 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.down_proj_weight_quantizers.118 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.35e-02 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.down_proj_weight_quantizers.119 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.down_proj_weight_quantizers.120 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.64e-02 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.down_proj_weight_quantizers.121 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.down_proj_weight_quantizers.122 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.30e-02 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.down_proj_weight_quantizers.123 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.down_proj_weight_quantizers.124 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.81e-01 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.down_proj_weight_quantizers.125 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.down_proj_weight_quantizers.126 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.down_proj_weight_quantizers.127 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.down_proj_weight_quantizers.128 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.down_proj_weight_quantizers.129 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.down_proj_weight_quantizers.130 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.27e-01 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.down_proj_weight_quantizers.131 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.down_proj_weight_quantizers.132 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.down_proj_weight_quantizers.133 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.46e-01 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.down_proj_weight_quantizers.134 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.down_proj_weight_quantizers.135 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.down_proj_weight_quantizers.136 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.64e-02 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.down_proj_weight_quantizers.137 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.down_proj_weight_quantizers.138 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.down_proj_weight_quantizers.139 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.down_proj_weight_quantizers.140 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.down_proj_weight_quantizers.141 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.down_proj_weight_quantizers.142 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.down_proj_weight_quantizers.143 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.down_proj_weight_quantizers.144 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.down_proj_weight_quantizers.145 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.down_proj_weight_quantizers.146 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.down_proj_weight_quantizers.147 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.down_proj_weight_quantizers.148 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.down_proj_weight_quantizers.149 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.55e-01 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.down_proj_weight_quantizers.150 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.down_proj_weight_quantizers.151 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.down_proj_weight_quantizers.152 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.32e-01 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.down_proj_weight_quantizers.153 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.59e-02 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.down_proj_weight_quantizers.154 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.down_proj_weight_quantizers.155 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.60e-01 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.down_proj_weight_quantizers.156 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.52e-01 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.down_proj_weight_quantizers.157 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.down_proj_weight_quantizers.158 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.down_proj_weight_quantizers.159 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.down_proj_weight_quantizers.160 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.down_proj_weight_quantizers.161 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.06e-01 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.down_proj_weight_quantizers.162 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.down_proj_weight_quantizers.163 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.down_proj_weight_quantizers.164 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.down_proj_weight_quantizers.165 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.down_proj_weight_quantizers.166 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.down_proj_weight_quantizers.167 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.down_proj_weight_quantizers.168 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.down_proj_weight_quantizers.169 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.down_proj_weight_quantizers.170 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.down_proj_weight_quantizers.171 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.down_proj_weight_quantizers.172 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.down_proj_weight_quantizers.173 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.down_proj_weight_quantizers.174 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.down_proj_weight_quantizers.175 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.down_proj_weight_quantizers.176 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.down_proj_weight_quantizers.177 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.down_proj_weight_quantizers.178 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.down_proj_weight_quantizers.179 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.down_proj_weight_quantizers.180 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.down_proj_weight_quantizers.181 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.59e-02 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.down_proj_weight_quantizers.182 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.down_proj_weight_quantizers.183 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.down_proj_weight_quantizers.184 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.down_proj_weight_quantizers.185 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.down_proj_weight_quantizers.186 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.down_proj_weight_quantizers.187 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.down_proj_weight_quantizers.188 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.64e-02 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.down_proj_weight_quantizers.189 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.down_proj_weight_quantizers.190 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.21e-01 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.down_proj_weight_quantizers.191 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.down_proj_weight_quantizers.192 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.50e-01 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.down_proj_weight_quantizers.193 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.59e-02 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.down_proj_weight_quantizers.194 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.64e-02 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.down_proj_weight_quantizers.195 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.down_proj_weight_quantizers.196 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.down_proj_weight_quantizers.197 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.65e-01 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.down_proj_weight_quantizers.198 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.down_proj_weight_quantizers.199 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.down_proj_weight_quantizers.200 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.down_proj_weight_quantizers.201 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.down_proj_weight_quantizers.202 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.64e-02 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.down_proj_weight_quantizers.203 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.down_proj_weight_quantizers.204 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.down_proj_weight_quantizers.205 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.14e-01 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.down_proj_weight_quantizers.206 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.down_proj_weight_quantizers.207 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.down_proj_weight_quantizers.208 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.down_proj_weight_quantizers.209 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.down_proj_weight_quantizers.210 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.down_proj_weight_quantizers.211 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.down_proj_weight_quantizers.212 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.down_proj_weight_quantizers.213 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.38e-01 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.down_proj_weight_quantizers.214 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.down_proj_weight_quantizers.215 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.down_proj_weight_quantizers.216 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.down_proj_weight_quantizers.217 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.64e-02 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.down_proj_weight_quantizers.218 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.down_proj_weight_quantizers.219 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.down_proj_weight_quantizers.220 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.down_proj_weight_quantizers.221 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.64e-02 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.down_proj_weight_quantizers.222 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.down_proj_weight_quantizers.223 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.down_proj_weight_quantizers.224 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.down_proj_weight_quantizers.225 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.down_proj_weight_quantizers.226 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.down_proj_weight_quantizers.227 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.down_proj_weight_quantizers.228 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.down_proj_weight_quantizers.229 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.63e-01 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.down_proj_weight_quantizers.230 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.down_proj_weight_quantizers.231 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.15e-01 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.down_proj_weight_quantizers.232 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.down_proj_weight_quantizers.233 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.down_proj_weight_quantizers.234 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.64e-02 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.down_proj_weight_quantizers.235 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.down_proj_weight_quantizers.236 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.down_proj_weight_quantizers.237 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.33e-01 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.down_proj_weight_quantizers.238 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.down_proj_weight_quantizers.239 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.down_proj_weight_quantizers.240 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.down_proj_weight_quantizers.241 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.down_proj_weight_quantizers.242 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.down_proj_weight_quantizers.243 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.37e-01 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.down_proj_weight_quantizers.244 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.down_proj_weight_quantizers.245 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.down_proj_weight_quantizers.246 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.down_proj_weight_quantizers.247 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.down_proj_weight_quantizers.248 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.down_proj_weight_quantizers.249 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.down_proj_weight_quantizers.250 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.down_proj_weight_quantizers.251 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.down_proj_weight_quantizers.252 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.65e-01 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.down_proj_weight_quantizers.253 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.21e-01 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.down_proj_weight_quantizers.254 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.down_proj_weight_quantizers.255 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.55.mlp.shared_experts.gate_proj.input_quantizer TensorQuantizer(disabled) +model.layers.55.mlp.shared_experts.gate_proj.output_quantizer TensorQuantizer(disabled) +model.layers.55.mlp.shared_experts.gate_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.55.mlp.shared_experts.up_proj.input_quantizer TensorQuantizer(disabled) +model.layers.55.mlp.shared_experts.up_proj.output_quantizer TensorQuantizer(disabled) +model.layers.55.mlp.shared_experts.up_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.55.mlp.shared_experts.down_proj.input_quantizer TensorQuantizer(disabled) +model.layers.55.mlp.shared_experts.down_proj.output_quantizer TensorQuantizer(disabled) +model.layers.55.mlp.shared_experts.down_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.56.self_attn.q_a_proj.input_quantizer TensorQuantizer(disabled) +model.layers.56.self_attn.q_a_proj.output_quantizer TensorQuantizer(disabled) +model.layers.56.self_attn.q_a_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.56.self_attn.q_b_proj.input_quantizer TensorQuantizer(disabled) +model.layers.56.self_attn.q_b_proj.output_quantizer TensorQuantizer(disabled) +model.layers.56.self_attn.q_b_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.56.self_attn.kv_a_proj_with_mqa.input_quantizer TensorQuantizer(disabled) +model.layers.56.self_attn.kv_a_proj_with_mqa.output_quantizer TensorQuantizer(disabled) +model.layers.56.self_attn.kv_a_proj_with_mqa.weight_quantizer TensorQuantizer(disabled) +model.layers.56.self_attn.kv_b_proj.input_quantizer TensorQuantizer(disabled) +model.layers.56.self_attn.kv_b_proj.output_quantizer TensorQuantizer(disabled) +model.layers.56.self_attn.kv_b_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.56.self_attn.o_proj.input_quantizer TensorQuantizer(disabled) +model.layers.56.self_attn.o_proj.output_quantizer TensorQuantizer(disabled) +model.layers.56.self_attn.o_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.56.self_attn.q_bmm_quantizer TensorQuantizer(disabled) +model.layers.56.self_attn.k_bmm_quantizer TensorQuantizer((4, 3) bit fake per-tensor amax=4.48e+02(const) calibrator=MaxCalibrator quant) +model.layers.56.self_attn.v_bmm_quantizer TensorQuantizer((4, 3) bit fake per-tensor amax=4.48e+02(const) calibrator=MaxCalibrator quant) +model.layers.56.self_attn.p_bmm_quantizer TensorQuantizer(disabled) +model.layers.56.mlp.experts.gate_up_proj_input_quantizer TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.69e+00 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.gate_up_proj_weight_quantizers.0 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.gate_up_proj_weight_quantizers.1 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.gate_up_proj_weight_quantizers.2 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.gate_up_proj_weight_quantizers.3 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.77e-01 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.gate_up_proj_weight_quantizers.4 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.gate_up_proj_weight_quantizers.5 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.gate_up_proj_weight_quantizers.6 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.gate_up_proj_weight_quantizers.7 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.gate_up_proj_weight_quantizers.8 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.gate_up_proj_weight_quantizers.9 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.79e-01 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.gate_up_proj_weight_quantizers.10 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.gate_up_proj_weight_quantizers.11 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.gate_up_proj_weight_quantizers.12 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.gate_up_proj_weight_quantizers.13 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.gate_up_proj_weight_quantizers.14 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.gate_up_proj_weight_quantizers.15 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.21e-01 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.gate_up_proj_weight_quantizers.16 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.gate_up_proj_weight_quantizers.17 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.gate_up_proj_weight_quantizers.18 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.gate_up_proj_weight_quantizers.19 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.gate_up_proj_weight_quantizers.20 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.gate_up_proj_weight_quantizers.21 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.gate_up_proj_weight_quantizers.22 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.gate_up_proj_weight_quantizers.23 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.gate_up_proj_weight_quantizers.24 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.gate_up_proj_weight_quantizers.25 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.gate_up_proj_weight_quantizers.26 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.gate_up_proj_weight_quantizers.27 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.gate_up_proj_weight_quantizers.28 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.gate_up_proj_weight_quantizers.29 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.gate_up_proj_weight_quantizers.30 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.55e-01 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.gate_up_proj_weight_quantizers.31 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.gate_up_proj_weight_quantizers.32 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.gate_up_proj_weight_quantizers.33 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.gate_up_proj_weight_quantizers.34 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.28e-01 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.gate_up_proj_weight_quantizers.35 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.72e-01 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.gate_up_proj_weight_quantizers.36 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.gate_up_proj_weight_quantizers.37 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.gate_up_proj_weight_quantizers.38 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.gate_up_proj_weight_quantizers.39 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.gate_up_proj_weight_quantizers.40 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.gate_up_proj_weight_quantizers.41 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.gate_up_proj_weight_quantizers.42 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.gate_up_proj_weight_quantizers.43 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.gate_up_proj_weight_quantizers.44 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.gate_up_proj_weight_quantizers.45 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.gate_up_proj_weight_quantizers.46 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.gate_up_proj_weight_quantizers.47 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.gate_up_proj_weight_quantizers.48 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.53e-01 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.gate_up_proj_weight_quantizers.49 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.gate_up_proj_weight_quantizers.50 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.gate_up_proj_weight_quantizers.51 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.gate_up_proj_weight_quantizers.52 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.gate_up_proj_weight_quantizers.53 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.gate_up_proj_weight_quantizers.54 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.gate_up_proj_weight_quantizers.55 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.gate_up_proj_weight_quantizers.56 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.gate_up_proj_weight_quantizers.57 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.20e-01 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.gate_up_proj_weight_quantizers.58 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.gate_up_proj_weight_quantizers.59 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.gate_up_proj_weight_quantizers.60 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.gate_up_proj_weight_quantizers.61 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.gate_up_proj_weight_quantizers.62 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.gate_up_proj_weight_quantizers.63 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.gate_up_proj_weight_quantizers.64 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.gate_up_proj_weight_quantizers.65 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.gate_up_proj_weight_quantizers.66 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.gate_up_proj_weight_quantizers.67 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.gate_up_proj_weight_quantizers.68 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.gate_up_proj_weight_quantizers.69 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.29e-01 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.gate_up_proj_weight_quantizers.70 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.gate_up_proj_weight_quantizers.71 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.gate_up_proj_weight_quantizers.72 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.gate_up_proj_weight_quantizers.73 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.gate_up_proj_weight_quantizers.74 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.gate_up_proj_weight_quantizers.75 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.gate_up_proj_weight_quantizers.76 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.gate_up_proj_weight_quantizers.77 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.24e-01 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.gate_up_proj_weight_quantizers.78 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.23e-01 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.gate_up_proj_weight_quantizers.79 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.13e-01 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.gate_up_proj_weight_quantizers.80 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.gate_up_proj_weight_quantizers.81 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.gate_up_proj_weight_quantizers.82 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.gate_up_proj_weight_quantizers.83 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.gate_up_proj_weight_quantizers.84 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.gate_up_proj_weight_quantizers.85 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.gate_up_proj_weight_quantizers.86 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.18e-01 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.gate_up_proj_weight_quantizers.87 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.gate_up_proj_weight_quantizers.88 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.gate_up_proj_weight_quantizers.89 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.gate_up_proj_weight_quantizers.90 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.05e-01 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.gate_up_proj_weight_quantizers.91 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.gate_up_proj_weight_quantizers.92 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.74e-01 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.gate_up_proj_weight_quantizers.93 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.16e-01 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.gate_up_proj_weight_quantizers.94 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.gate_up_proj_weight_quantizers.95 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.gate_up_proj_weight_quantizers.96 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.42e-01 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.gate_up_proj_weight_quantizers.97 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.gate_up_proj_weight_quantizers.98 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.18e-01 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.gate_up_proj_weight_quantizers.99 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.gate_up_proj_weight_quantizers.100 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.gate_up_proj_weight_quantizers.101 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.84e-01 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.gate_up_proj_weight_quantizers.102 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.gate_up_proj_weight_quantizers.103 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.14e-01 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.gate_up_proj_weight_quantizers.104 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.gate_up_proj_weight_quantizers.105 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.gate_up_proj_weight_quantizers.106 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.27e-01 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.gate_up_proj_weight_quantizers.107 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.48e-01 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.gate_up_proj_weight_quantizers.108 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.37e-01 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.gate_up_proj_weight_quantizers.109 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.gate_up_proj_weight_quantizers.110 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.gate_up_proj_weight_quantizers.111 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.gate_up_proj_weight_quantizers.112 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.gate_up_proj_weight_quantizers.113 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.gate_up_proj_weight_quantizers.114 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.85e-01 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.gate_up_proj_weight_quantizers.115 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.gate_up_proj_weight_quantizers.116 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.gate_up_proj_weight_quantizers.117 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.gate_up_proj_weight_quantizers.118 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.gate_up_proj_weight_quantizers.119 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.gate_up_proj_weight_quantizers.120 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.gate_up_proj_weight_quantizers.121 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.17e-01 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.gate_up_proj_weight_quantizers.122 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.70e-01 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.gate_up_proj_weight_quantizers.123 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.gate_up_proj_weight_quantizers.124 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.gate_up_proj_weight_quantizers.125 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.gate_up_proj_weight_quantizers.126 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.83e-01 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.gate_up_proj_weight_quantizers.127 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.gate_up_proj_weight_quantizers.128 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.gate_up_proj_weight_quantizers.129 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.54e-01 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.gate_up_proj_weight_quantizers.130 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.gate_up_proj_weight_quantizers.131 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.72e-01 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.gate_up_proj_weight_quantizers.132 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.gate_up_proj_weight_quantizers.133 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.gate_up_proj_weight_quantizers.134 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.15e-01 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.gate_up_proj_weight_quantizers.135 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.gate_up_proj_weight_quantizers.136 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.gate_up_proj_weight_quantizers.137 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.gate_up_proj_weight_quantizers.138 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.gate_up_proj_weight_quantizers.139 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.gate_up_proj_weight_quantizers.140 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.gate_up_proj_weight_quantizers.141 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.03e-01 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.gate_up_proj_weight_quantizers.142 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.gate_up_proj_weight_quantizers.143 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.gate_up_proj_weight_quantizers.144 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.gate_up_proj_weight_quantizers.145 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.gate_up_proj_weight_quantizers.146 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.gate_up_proj_weight_quantizers.147 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.20e-01 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.gate_up_proj_weight_quantizers.148 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.05e-01 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.gate_up_proj_weight_quantizers.149 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.gate_up_proj_weight_quantizers.150 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.gate_up_proj_weight_quantizers.151 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.gate_up_proj_weight_quantizers.152 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.gate_up_proj_weight_quantizers.153 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.gate_up_proj_weight_quantizers.154 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.gate_up_proj_weight_quantizers.155 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.26e-01 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.gate_up_proj_weight_quantizers.156 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.gate_up_proj_weight_quantizers.157 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.68e-01 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.gate_up_proj_weight_quantizers.158 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.gate_up_proj_weight_quantizers.159 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.gate_up_proj_weight_quantizers.160 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.48e-01 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.gate_up_proj_weight_quantizers.161 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.gate_up_proj_weight_quantizers.162 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.gate_up_proj_weight_quantizers.163 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.gate_up_proj_weight_quantizers.164 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.gate_up_proj_weight_quantizers.165 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.gate_up_proj_weight_quantizers.166 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.gate_up_proj_weight_quantizers.167 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.gate_up_proj_weight_quantizers.168 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.gate_up_proj_weight_quantizers.169 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.gate_up_proj_weight_quantizers.170 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.gate_up_proj_weight_quantizers.171 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.gate_up_proj_weight_quantizers.172 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.46e-01 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.gate_up_proj_weight_quantizers.173 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.gate_up_proj_weight_quantizers.174 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.gate_up_proj_weight_quantizers.175 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.25e-01 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.gate_up_proj_weight_quantizers.176 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.gate_up_proj_weight_quantizers.177 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.gate_up_proj_weight_quantizers.178 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.13e-01 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.gate_up_proj_weight_quantizers.179 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.gate_up_proj_weight_quantizers.180 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.gate_up_proj_weight_quantizers.181 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.gate_up_proj_weight_quantizers.182 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.25e-01 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.gate_up_proj_weight_quantizers.183 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.gate_up_proj_weight_quantizers.184 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.gate_up_proj_weight_quantizers.185 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.gate_up_proj_weight_quantizers.186 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.17e-01 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.gate_up_proj_weight_quantizers.187 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.gate_up_proj_weight_quantizers.188 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.68e-01 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.gate_up_proj_weight_quantizers.189 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.gate_up_proj_weight_quantizers.190 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.gate_up_proj_weight_quantizers.191 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.gate_up_proj_weight_quantizers.192 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.gate_up_proj_weight_quantizers.193 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.gate_up_proj_weight_quantizers.194 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.gate_up_proj_weight_quantizers.195 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.gate_up_proj_weight_quantizers.196 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.26e-01 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.gate_up_proj_weight_quantizers.197 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.gate_up_proj_weight_quantizers.198 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.gate_up_proj_weight_quantizers.199 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.gate_up_proj_weight_quantizers.200 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.gate_up_proj_weight_quantizers.201 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.09e-01 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.gate_up_proj_weight_quantizers.202 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.gate_up_proj_weight_quantizers.203 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.gate_up_proj_weight_quantizers.204 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.gate_up_proj_weight_quantizers.205 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.gate_up_proj_weight_quantizers.206 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.gate_up_proj_weight_quantizers.207 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.gate_up_proj_weight_quantizers.208 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.gate_up_proj_weight_quantizers.209 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.gate_up_proj_weight_quantizers.210 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.gate_up_proj_weight_quantizers.211 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.gate_up_proj_weight_quantizers.212 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.gate_up_proj_weight_quantizers.213 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.gate_up_proj_weight_quantizers.214 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.gate_up_proj_weight_quantizers.215 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.gate_up_proj_weight_quantizers.216 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.13e-01 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.gate_up_proj_weight_quantizers.217 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.gate_up_proj_weight_quantizers.218 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.gate_up_proj_weight_quantizers.219 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.gate_up_proj_weight_quantizers.220 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.gate_up_proj_weight_quantizers.221 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.gate_up_proj_weight_quantizers.222 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.gate_up_proj_weight_quantizers.223 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.gate_up_proj_weight_quantizers.224 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.29e-01 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.gate_up_proj_weight_quantizers.225 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.gate_up_proj_weight_quantizers.226 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.gate_up_proj_weight_quantizers.227 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.15e-01 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.gate_up_proj_weight_quantizers.228 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.gate_up_proj_weight_quantizers.229 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.gate_up_proj_weight_quantizers.230 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.gate_up_proj_weight_quantizers.231 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.gate_up_proj_weight_quantizers.232 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.gate_up_proj_weight_quantizers.233 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.gate_up_proj_weight_quantizers.234 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.gate_up_proj_weight_quantizers.235 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.73e-01 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.gate_up_proj_weight_quantizers.236 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.gate_up_proj_weight_quantizers.237 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.gate_up_proj_weight_quantizers.238 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.gate_up_proj_weight_quantizers.239 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.gate_up_proj_weight_quantizers.240 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.gate_up_proj_weight_quantizers.241 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.gate_up_proj_weight_quantizers.242 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.gate_up_proj_weight_quantizers.243 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.85e-01 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.gate_up_proj_weight_quantizers.244 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.gate_up_proj_weight_quantizers.245 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.gate_up_proj_weight_quantizers.246 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.gate_up_proj_weight_quantizers.247 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.gate_up_proj_weight_quantizers.248 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.47e-01 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.gate_up_proj_weight_quantizers.249 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.gate_up_proj_weight_quantizers.250 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.gate_up_proj_weight_quantizers.251 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=5.23e-01 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.gate_up_proj_weight_quantizers.252 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.gate_up_proj_weight_quantizers.253 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.54e-01 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.gate_up_proj_weight_quantizers.254 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.gate_up_proj_weight_quantizers.255 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.down_proj_input_quantizer TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.55e+01 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.down_proj_weight_quantizers.0 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.down_proj_weight_quantizers.1 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.down_proj_weight_quantizers.2 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.down_proj_weight_quantizers.3 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.20e-01 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.down_proj_weight_quantizers.4 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.down_proj_weight_quantizers.5 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.down_proj_weight_quantizers.6 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.down_proj_weight_quantizers.7 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.down_proj_weight_quantizers.8 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.down_proj_weight_quantizers.9 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.down_proj_weight_quantizers.10 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.down_proj_weight_quantizers.11 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.down_proj_weight_quantizers.12 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.down_proj_weight_quantizers.13 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.down_proj_weight_quantizers.14 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.down_proj_weight_quantizers.15 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.19e-01 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.down_proj_weight_quantizers.16 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.down_proj_weight_quantizers.17 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.down_proj_weight_quantizers.18 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.down_proj_weight_quantizers.19 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.17e-01 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.down_proj_weight_quantizers.20 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.64e-02 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.down_proj_weight_quantizers.21 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.down_proj_weight_quantizers.22 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.down_proj_weight_quantizers.23 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.40e-02 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.down_proj_weight_quantizers.24 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.50e-01 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.down_proj_weight_quantizers.25 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.down_proj_weight_quantizers.26 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.down_proj_weight_quantizers.27 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.down_proj_weight_quantizers.28 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.down_proj_weight_quantizers.29 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.down_proj_weight_quantizers.30 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.down_proj_weight_quantizers.31 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.down_proj_weight_quantizers.32 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.down_proj_weight_quantizers.33 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.down_proj_weight_quantizers.34 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.down_proj_weight_quantizers.35 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.52e-01 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.down_proj_weight_quantizers.36 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.down_proj_weight_quantizers.37 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.down_proj_weight_quantizers.38 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.down_proj_weight_quantizers.39 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.down_proj_weight_quantizers.40 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.down_proj_weight_quantizers.41 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.down_proj_weight_quantizers.42 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.down_proj_weight_quantizers.43 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.down_proj_weight_quantizers.44 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.down_proj_weight_quantizers.45 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.down_proj_weight_quantizers.46 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.down_proj_weight_quantizers.47 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.down_proj_weight_quantizers.48 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.down_proj_weight_quantizers.49 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.down_proj_weight_quantizers.50 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.50e-02 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.down_proj_weight_quantizers.51 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.down_proj_weight_quantizers.52 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.down_proj_weight_quantizers.53 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.down_proj_weight_quantizers.54 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.down_proj_weight_quantizers.55 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.down_proj_weight_quantizers.56 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.down_proj_weight_quantizers.57 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.down_proj_weight_quantizers.58 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.23e-01 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.down_proj_weight_quantizers.59 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.down_proj_weight_quantizers.60 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.down_proj_weight_quantizers.61 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.down_proj_weight_quantizers.62 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.down_proj_weight_quantizers.63 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.down_proj_weight_quantizers.64 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.down_proj_weight_quantizers.65 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.down_proj_weight_quantizers.66 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.down_proj_weight_quantizers.67 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.down_proj_weight_quantizers.68 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.down_proj_weight_quantizers.69 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.down_proj_weight_quantizers.70 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.down_proj_weight_quantizers.71 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.down_proj_weight_quantizers.72 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.down_proj_weight_quantizers.73 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.down_proj_weight_quantizers.74 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.down_proj_weight_quantizers.75 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.down_proj_weight_quantizers.76 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.down_proj_weight_quantizers.77 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.65e-01 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.down_proj_weight_quantizers.78 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.down_proj_weight_quantizers.79 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.down_proj_weight_quantizers.80 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.down_proj_weight_quantizers.81 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.down_proj_weight_quantizers.82 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.down_proj_weight_quantizers.83 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.down_proj_weight_quantizers.84 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.down_proj_weight_quantizers.85 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.down_proj_weight_quantizers.86 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.down_proj_weight_quantizers.87 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.down_proj_weight_quantizers.88 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.down_proj_weight_quantizers.89 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.down_proj_weight_quantizers.90 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.down_proj_weight_quantizers.91 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.down_proj_weight_quantizers.92 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.down_proj_weight_quantizers.93 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.down_proj_weight_quantizers.94 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.down_proj_weight_quantizers.95 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.down_proj_weight_quantizers.96 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.28e-01 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.down_proj_weight_quantizers.97 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.down_proj_weight_quantizers.98 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.35e-01 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.down_proj_weight_quantizers.99 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.down_proj_weight_quantizers.100 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.down_proj_weight_quantizers.101 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.down_proj_weight_quantizers.102 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.down_proj_weight_quantizers.103 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.down_proj_weight_quantizers.104 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.down_proj_weight_quantizers.105 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.06e-01 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.down_proj_weight_quantizers.106 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.down_proj_weight_quantizers.107 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.23e-01 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.down_proj_weight_quantizers.108 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.down_proj_weight_quantizers.109 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.down_proj_weight_quantizers.110 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.down_proj_weight_quantizers.111 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.down_proj_weight_quantizers.112 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.down_proj_weight_quantizers.113 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.down_proj_weight_quantizers.114 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.down_proj_weight_quantizers.115 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.down_proj_weight_quantizers.116 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.down_proj_weight_quantizers.117 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.down_proj_weight_quantizers.118 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.down_proj_weight_quantizers.119 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.down_proj_weight_quantizers.120 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.down_proj_weight_quantizers.121 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.down_proj_weight_quantizers.122 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.down_proj_weight_quantizers.123 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.down_proj_weight_quantizers.124 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.50e-02 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.down_proj_weight_quantizers.125 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.down_proj_weight_quantizers.126 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.down_proj_weight_quantizers.127 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.down_proj_weight_quantizers.128 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.down_proj_weight_quantizers.129 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.52e-01 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.down_proj_weight_quantizers.130 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.59e-02 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.down_proj_weight_quantizers.131 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.down_proj_weight_quantizers.132 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.down_proj_weight_quantizers.133 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.down_proj_weight_quantizers.134 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.down_proj_weight_quantizers.135 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.down_proj_weight_quantizers.136 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.down_proj_weight_quantizers.137 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.down_proj_weight_quantizers.138 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.down_proj_weight_quantizers.139 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.down_proj_weight_quantizers.140 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.down_proj_weight_quantizers.141 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.79e-01 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.down_proj_weight_quantizers.142 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.down_proj_weight_quantizers.143 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.down_proj_weight_quantizers.144 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.59e-02 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.down_proj_weight_quantizers.145 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.down_proj_weight_quantizers.146 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.down_proj_weight_quantizers.147 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.down_proj_weight_quantizers.148 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.46e-01 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.down_proj_weight_quantizers.149 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.down_proj_weight_quantizers.150 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.down_proj_weight_quantizers.151 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.down_proj_weight_quantizers.152 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.down_proj_weight_quantizers.153 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.16e-01 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.down_proj_weight_quantizers.154 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.down_proj_weight_quantizers.155 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.79e-01 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.down_proj_weight_quantizers.156 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.down_proj_weight_quantizers.157 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.down_proj_weight_quantizers.158 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.down_proj_weight_quantizers.159 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.down_proj_weight_quantizers.160 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.down_proj_weight_quantizers.161 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.down_proj_weight_quantizers.162 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.down_proj_weight_quantizers.163 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.down_proj_weight_quantizers.164 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.down_proj_weight_quantizers.165 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.down_proj_weight_quantizers.166 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.27e-01 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.down_proj_weight_quantizers.167 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.45e-02 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.down_proj_weight_quantizers.168 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.down_proj_weight_quantizers.169 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.59e-02 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.down_proj_weight_quantizers.170 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.down_proj_weight_quantizers.171 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.down_proj_weight_quantizers.172 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.26e-01 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.down_proj_weight_quantizers.173 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.down_proj_weight_quantizers.174 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.down_proj_weight_quantizers.175 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.down_proj_weight_quantizers.176 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.18e-01 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.down_proj_weight_quantizers.177 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.50e-02 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.down_proj_weight_quantizers.178 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.down_proj_weight_quantizers.179 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.down_proj_weight_quantizers.180 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.down_proj_weight_quantizers.181 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.down_proj_weight_quantizers.182 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.down_proj_weight_quantizers.183 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.down_proj_weight_quantizers.184 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.down_proj_weight_quantizers.185 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.down_proj_weight_quantizers.186 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.down_proj_weight_quantizers.187 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.down_proj_weight_quantizers.188 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.down_proj_weight_quantizers.189 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.down_proj_weight_quantizers.190 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.down_proj_weight_quantizers.191 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.down_proj_weight_quantizers.192 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.down_proj_weight_quantizers.193 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.down_proj_weight_quantizers.194 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.down_proj_weight_quantizers.195 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.down_proj_weight_quantizers.196 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.down_proj_weight_quantizers.197 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.down_proj_weight_quantizers.198 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.down_proj_weight_quantizers.199 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.59e-02 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.down_proj_weight_quantizers.200 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.down_proj_weight_quantizers.201 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.46e-01 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.down_proj_weight_quantizers.202 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.down_proj_weight_quantizers.203 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.down_proj_weight_quantizers.204 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.down_proj_weight_quantizers.205 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.down_proj_weight_quantizers.206 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.down_proj_weight_quantizers.207 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.down_proj_weight_quantizers.208 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.down_proj_weight_quantizers.209 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.down_proj_weight_quantizers.210 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.down_proj_weight_quantizers.211 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.down_proj_weight_quantizers.212 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.down_proj_weight_quantizers.213 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.down_proj_weight_quantizers.214 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.64e-02 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.down_proj_weight_quantizers.215 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.down_proj_weight_quantizers.216 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.down_proj_weight_quantizers.217 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.down_proj_weight_quantizers.218 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.down_proj_weight_quantizers.219 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.down_proj_weight_quantizers.220 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.59e-01 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.down_proj_weight_quantizers.221 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.down_proj_weight_quantizers.222 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.down_proj_weight_quantizers.223 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.down_proj_weight_quantizers.224 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.down_proj_weight_quantizers.225 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.35e-01 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.down_proj_weight_quantizers.226 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.down_proj_weight_quantizers.227 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.down_proj_weight_quantizers.228 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.down_proj_weight_quantizers.229 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.down_proj_weight_quantizers.230 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.down_proj_weight_quantizers.231 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.59e-02 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.down_proj_weight_quantizers.232 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.down_proj_weight_quantizers.233 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.down_proj_weight_quantizers.234 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.down_proj_weight_quantizers.235 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.down_proj_weight_quantizers.236 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.down_proj_weight_quantizers.237 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.down_proj_weight_quantizers.238 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.64e-02 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.down_proj_weight_quantizers.239 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.35e-02 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.down_proj_weight_quantizers.240 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.89e-01 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.down_proj_weight_quantizers.241 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.down_proj_weight_quantizers.242 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.down_proj_weight_quantizers.243 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.71e-01 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.down_proj_weight_quantizers.244 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.down_proj_weight_quantizers.245 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.down_proj_weight_quantizers.246 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.down_proj_weight_quantizers.247 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.down_proj_weight_quantizers.248 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.78e-01 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.down_proj_weight_quantizers.249 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.down_proj_weight_quantizers.250 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.down_proj_weight_quantizers.251 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.44e-01 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.down_proj_weight_quantizers.252 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.down_proj_weight_quantizers.253 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.52e-01 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.down_proj_weight_quantizers.254 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.down_proj_weight_quantizers.255 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.64e-02 calibrator=MaxCalibrator quant) +model.layers.56.mlp.shared_experts.gate_proj.input_quantizer TensorQuantizer(disabled) +model.layers.56.mlp.shared_experts.gate_proj.output_quantizer TensorQuantizer(disabled) +model.layers.56.mlp.shared_experts.gate_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.56.mlp.shared_experts.up_proj.input_quantizer TensorQuantizer(disabled) +model.layers.56.mlp.shared_experts.up_proj.output_quantizer TensorQuantizer(disabled) +model.layers.56.mlp.shared_experts.up_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.56.mlp.shared_experts.down_proj.input_quantizer TensorQuantizer(disabled) +model.layers.56.mlp.shared_experts.down_proj.output_quantizer TensorQuantizer(disabled) +model.layers.56.mlp.shared_experts.down_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.57.self_attn.q_a_proj.input_quantizer TensorQuantizer(disabled) +model.layers.57.self_attn.q_a_proj.output_quantizer TensorQuantizer(disabled) +model.layers.57.self_attn.q_a_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.57.self_attn.q_b_proj.input_quantizer TensorQuantizer(disabled) +model.layers.57.self_attn.q_b_proj.output_quantizer TensorQuantizer(disabled) +model.layers.57.self_attn.q_b_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.57.self_attn.kv_a_proj_with_mqa.input_quantizer TensorQuantizer(disabled) +model.layers.57.self_attn.kv_a_proj_with_mqa.output_quantizer TensorQuantizer(disabled) +model.layers.57.self_attn.kv_a_proj_with_mqa.weight_quantizer TensorQuantizer(disabled) +model.layers.57.self_attn.kv_b_proj.input_quantizer TensorQuantizer(disabled) +model.layers.57.self_attn.kv_b_proj.output_quantizer TensorQuantizer(disabled) +model.layers.57.self_attn.kv_b_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.57.self_attn.o_proj.input_quantizer TensorQuantizer(disabled) +model.layers.57.self_attn.o_proj.output_quantizer TensorQuantizer(disabled) +model.layers.57.self_attn.o_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.57.self_attn.q_bmm_quantizer TensorQuantizer(disabled) +model.layers.57.self_attn.k_bmm_quantizer TensorQuantizer((4, 3) bit fake per-tensor amax=4.48e+02(const) calibrator=MaxCalibrator quant) +model.layers.57.self_attn.v_bmm_quantizer TensorQuantizer((4, 3) bit fake per-tensor amax=4.48e+02(const) calibrator=MaxCalibrator quant) +model.layers.57.self_attn.p_bmm_quantizer TensorQuantizer(disabled) +model.layers.57.mlp.experts.gate_up_proj_input_quantizer TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.75e+00 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.gate_up_proj_weight_quantizers.0 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.gate_up_proj_weight_quantizers.1 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.24e-01 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.gate_up_proj_weight_quantizers.2 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.gate_up_proj_weight_quantizers.3 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.gate_up_proj_weight_quantizers.4 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.gate_up_proj_weight_quantizers.5 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.gate_up_proj_weight_quantizers.6 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.gate_up_proj_weight_quantizers.7 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.gate_up_proj_weight_quantizers.8 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.38e-01 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.gate_up_proj_weight_quantizers.9 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.gate_up_proj_weight_quantizers.10 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.gate_up_proj_weight_quantizers.11 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.29e-01 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.gate_up_proj_weight_quantizers.12 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.gate_up_proj_weight_quantizers.13 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.gate_up_proj_weight_quantizers.14 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.36e-01 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.gate_up_proj_weight_quantizers.15 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.gate_up_proj_weight_quantizers.16 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.23e-01 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.gate_up_proj_weight_quantizers.17 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.gate_up_proj_weight_quantizers.18 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.gate_up_proj_weight_quantizers.19 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.14e-01 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.gate_up_proj_weight_quantizers.20 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.09e-01 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.gate_up_proj_weight_quantizers.21 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.gate_up_proj_weight_quantizers.22 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.gate_up_proj_weight_quantizers.23 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.28e-01 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.gate_up_proj_weight_quantizers.24 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.gate_up_proj_weight_quantizers.25 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.gate_up_proj_weight_quantizers.26 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.gate_up_proj_weight_quantizers.27 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.18e-01 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.gate_up_proj_weight_quantizers.28 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.gate_up_proj_weight_quantizers.29 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.gate_up_proj_weight_quantizers.30 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.gate_up_proj_weight_quantizers.31 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.gate_up_proj_weight_quantizers.32 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.14e-01 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.gate_up_proj_weight_quantizers.33 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.gate_up_proj_weight_quantizers.34 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.gate_up_proj_weight_quantizers.35 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.68e-01 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.gate_up_proj_weight_quantizers.36 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.19e-01 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.gate_up_proj_weight_quantizers.37 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.gate_up_proj_weight_quantizers.38 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.gate_up_proj_weight_quantizers.39 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.gate_up_proj_weight_quantizers.40 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.13e-01 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.gate_up_proj_weight_quantizers.41 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.gate_up_proj_weight_quantizers.42 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.gate_up_proj_weight_quantizers.43 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.34e-01 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.gate_up_proj_weight_quantizers.44 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.25e-01 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.gate_up_proj_weight_quantizers.45 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.16e-01 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.gate_up_proj_weight_quantizers.46 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.gate_up_proj_weight_quantizers.47 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.gate_up_proj_weight_quantizers.48 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.gate_up_proj_weight_quantizers.49 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.gate_up_proj_weight_quantizers.50 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.gate_up_proj_weight_quantizers.51 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.gate_up_proj_weight_quantizers.52 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.gate_up_proj_weight_quantizers.53 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.gate_up_proj_weight_quantizers.54 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.gate_up_proj_weight_quantizers.55 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.gate_up_proj_weight_quantizers.56 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.gate_up_proj_weight_quantizers.57 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.gate_up_proj_weight_quantizers.58 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.gate_up_proj_weight_quantizers.59 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.gate_up_proj_weight_quantizers.60 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.gate_up_proj_weight_quantizers.61 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.gate_up_proj_weight_quantizers.62 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.gate_up_proj_weight_quantizers.63 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.gate_up_proj_weight_quantizers.64 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.19e-01 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.gate_up_proj_weight_quantizers.65 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.gate_up_proj_weight_quantizers.66 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.22e-01 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.gate_up_proj_weight_quantizers.67 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.gate_up_proj_weight_quantizers.68 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.gate_up_proj_weight_quantizers.69 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.gate_up_proj_weight_quantizers.70 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.gate_up_proj_weight_quantizers.71 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.gate_up_proj_weight_quantizers.72 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.gate_up_proj_weight_quantizers.73 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.gate_up_proj_weight_quantizers.74 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.gate_up_proj_weight_quantizers.75 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.gate_up_proj_weight_quantizers.76 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.gate_up_proj_weight_quantizers.77 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.gate_up_proj_weight_quantizers.78 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.gate_up_proj_weight_quantizers.79 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.gate_up_proj_weight_quantizers.80 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.gate_up_proj_weight_quantizers.81 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.25e-01 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.gate_up_proj_weight_quantizers.82 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.gate_up_proj_weight_quantizers.83 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.gate_up_proj_weight_quantizers.84 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.gate_up_proj_weight_quantizers.85 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.20e-01 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.gate_up_proj_weight_quantizers.86 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.39e-01 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.gate_up_proj_weight_quantizers.87 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.gate_up_proj_weight_quantizers.88 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.gate_up_proj_weight_quantizers.89 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.gate_up_proj_weight_quantizers.90 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.gate_up_proj_weight_quantizers.91 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.gate_up_proj_weight_quantizers.92 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.gate_up_proj_weight_quantizers.93 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.gate_up_proj_weight_quantizers.94 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.gate_up_proj_weight_quantizers.95 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.gate_up_proj_weight_quantizers.96 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.gate_up_proj_weight_quantizers.97 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.gate_up_proj_weight_quantizers.98 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.gate_up_proj_weight_quantizers.99 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.gate_up_proj_weight_quantizers.100 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.gate_up_proj_weight_quantizers.101 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.gate_up_proj_weight_quantizers.102 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.29e-01 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.gate_up_proj_weight_quantizers.103 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=4.24e-01 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.gate_up_proj_weight_quantizers.104 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.gate_up_proj_weight_quantizers.105 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.55e-01 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.gate_up_proj_weight_quantizers.106 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.23e-01 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.gate_up_proj_weight_quantizers.107 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.16e-01 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.gate_up_proj_weight_quantizers.108 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.gate_up_proj_weight_quantizers.109 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.gate_up_proj_weight_quantizers.110 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.gate_up_proj_weight_quantizers.111 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.gate_up_proj_weight_quantizers.112 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.gate_up_proj_weight_quantizers.113 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.gate_up_proj_weight_quantizers.114 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.gate_up_proj_weight_quantizers.115 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.gate_up_proj_weight_quantizers.116 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.gate_up_proj_weight_quantizers.117 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.gate_up_proj_weight_quantizers.118 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.gate_up_proj_weight_quantizers.119 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.gate_up_proj_weight_quantizers.120 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.gate_up_proj_weight_quantizers.121 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.gate_up_proj_weight_quantizers.122 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.29e-01 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.gate_up_proj_weight_quantizers.123 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.gate_up_proj_weight_quantizers.124 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.gate_up_proj_weight_quantizers.125 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.13e-01 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.gate_up_proj_weight_quantizers.126 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.gate_up_proj_weight_quantizers.127 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.gate_up_proj_weight_quantizers.128 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.gate_up_proj_weight_quantizers.129 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.gate_up_proj_weight_quantizers.130 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.42e-01 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.gate_up_proj_weight_quantizers.131 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.43e-01 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.gate_up_proj_weight_quantizers.132 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.56e-01 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.gate_up_proj_weight_quantizers.133 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.gate_up_proj_weight_quantizers.134 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.21e-01 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.gate_up_proj_weight_quantizers.135 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.gate_up_proj_weight_quantizers.136 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.gate_up_proj_weight_quantizers.137 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.gate_up_proj_weight_quantizers.138 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.gate_up_proj_weight_quantizers.139 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.29e-01 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.gate_up_proj_weight_quantizers.140 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.gate_up_proj_weight_quantizers.141 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.95e-01 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.gate_up_proj_weight_quantizers.142 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.gate_up_proj_weight_quantizers.143 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.23e-01 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.gate_up_proj_weight_quantizers.144 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.gate_up_proj_weight_quantizers.145 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.16e-01 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.gate_up_proj_weight_quantizers.146 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.gate_up_proj_weight_quantizers.147 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.gate_up_proj_weight_quantizers.148 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.gate_up_proj_weight_quantizers.149 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.gate_up_proj_weight_quantizers.150 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.gate_up_proj_weight_quantizers.151 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.gate_up_proj_weight_quantizers.152 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.gate_up_proj_weight_quantizers.153 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.gate_up_proj_weight_quantizers.154 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.gate_up_proj_weight_quantizers.155 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.16e-01 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.gate_up_proj_weight_quantizers.156 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.gate_up_proj_weight_quantizers.157 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.37e-01 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.gate_up_proj_weight_quantizers.158 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.05e-01 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.gate_up_proj_weight_quantizers.159 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.gate_up_proj_weight_quantizers.160 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.gate_up_proj_weight_quantizers.161 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.gate_up_proj_weight_quantizers.162 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.13e-01 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.gate_up_proj_weight_quantizers.163 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.gate_up_proj_weight_quantizers.164 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.gate_up_proj_weight_quantizers.165 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.gate_up_proj_weight_quantizers.166 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.33e-01 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.gate_up_proj_weight_quantizers.167 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.gate_up_proj_weight_quantizers.168 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.gate_up_proj_weight_quantizers.169 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.gate_up_proj_weight_quantizers.170 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.gate_up_proj_weight_quantizers.171 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.gate_up_proj_weight_quantizers.172 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.24e-01 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.gate_up_proj_weight_quantizers.173 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.gate_up_proj_weight_quantizers.174 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.gate_up_proj_weight_quantizers.175 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.64e-01 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.gate_up_proj_weight_quantizers.176 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.gate_up_proj_weight_quantizers.177 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.45e-01 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.gate_up_proj_weight_quantizers.178 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.gate_up_proj_weight_quantizers.179 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.gate_up_proj_weight_quantizers.180 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.gate_up_proj_weight_quantizers.181 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.gate_up_proj_weight_quantizers.182 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.16e-01 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.gate_up_proj_weight_quantizers.183 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.gate_up_proj_weight_quantizers.184 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.gate_up_proj_weight_quantizers.185 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.18e-01 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.gate_up_proj_weight_quantizers.186 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.gate_up_proj_weight_quantizers.187 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.47e-01 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.gate_up_proj_weight_quantizers.188 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.gate_up_proj_weight_quantizers.189 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.gate_up_proj_weight_quantizers.190 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.51e-01 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.gate_up_proj_weight_quantizers.191 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.gate_up_proj_weight_quantizers.192 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.gate_up_proj_weight_quantizers.193 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.gate_up_proj_weight_quantizers.194 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.gate_up_proj_weight_quantizers.195 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.gate_up_proj_weight_quantizers.196 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.gate_up_proj_weight_quantizers.197 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.31e-01 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.gate_up_proj_weight_quantizers.198 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.gate_up_proj_weight_quantizers.199 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.14e-01 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.gate_up_proj_weight_quantizers.200 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.gate_up_proj_weight_quantizers.201 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.gate_up_proj_weight_quantizers.202 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.gate_up_proj_weight_quantizers.203 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.gate_up_proj_weight_quantizers.204 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.gate_up_proj_weight_quantizers.205 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.gate_up_proj_weight_quantizers.206 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.gate_up_proj_weight_quantizers.207 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.gate_up_proj_weight_quantizers.208 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.gate_up_proj_weight_quantizers.209 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.gate_up_proj_weight_quantizers.210 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.gate_up_proj_weight_quantizers.211 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.gate_up_proj_weight_quantizers.212 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.gate_up_proj_weight_quantizers.213 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.gate_up_proj_weight_quantizers.214 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.gate_up_proj_weight_quantizers.215 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.gate_up_proj_weight_quantizers.216 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.gate_up_proj_weight_quantizers.217 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.gate_up_proj_weight_quantizers.218 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.gate_up_proj_weight_quantizers.219 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.gate_up_proj_weight_quantizers.220 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.15e-01 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.gate_up_proj_weight_quantizers.221 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.gate_up_proj_weight_quantizers.222 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.16e-01 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.gate_up_proj_weight_quantizers.223 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.gate_up_proj_weight_quantizers.224 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.gate_up_proj_weight_quantizers.225 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.gate_up_proj_weight_quantizers.226 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.87e-01 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.gate_up_proj_weight_quantizers.227 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.gate_up_proj_weight_quantizers.228 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.gate_up_proj_weight_quantizers.229 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.gate_up_proj_weight_quantizers.230 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.gate_up_proj_weight_quantizers.231 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.gate_up_proj_weight_quantizers.232 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.gate_up_proj_weight_quantizers.233 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.gate_up_proj_weight_quantizers.234 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.26e-01 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.gate_up_proj_weight_quantizers.235 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.gate_up_proj_weight_quantizers.236 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.gate_up_proj_weight_quantizers.237 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.81e-01 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.gate_up_proj_weight_quantizers.238 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.gate_up_proj_weight_quantizers.239 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.gate_up_proj_weight_quantizers.240 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.24e-01 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.gate_up_proj_weight_quantizers.241 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.31e-01 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.gate_up_proj_weight_quantizers.242 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=5.00e-01 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.gate_up_proj_weight_quantizers.243 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.gate_up_proj_weight_quantizers.244 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.42e-01 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.gate_up_proj_weight_quantizers.245 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.gate_up_proj_weight_quantizers.246 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.gate_up_proj_weight_quantizers.247 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.gate_up_proj_weight_quantizers.248 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.gate_up_proj_weight_quantizers.249 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.gate_up_proj_weight_quantizers.250 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.gate_up_proj_weight_quantizers.251 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.gate_up_proj_weight_quantizers.252 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.gate_up_proj_weight_quantizers.253 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.gate_up_proj_weight_quantizers.254 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.gate_up_proj_weight_quantizers.255 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.down_proj_input_quantizer TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=4.12e+01 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.down_proj_weight_quantizers.0 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.down_proj_weight_quantizers.1 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.down_proj_weight_quantizers.2 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.27e-01 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.down_proj_weight_quantizers.3 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.down_proj_weight_quantizers.4 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.down_proj_weight_quantizers.5 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.down_proj_weight_quantizers.6 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.down_proj_weight_quantizers.7 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.down_proj_weight_quantizers.8 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.down_proj_weight_quantizers.9 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.down_proj_weight_quantizers.10 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.down_proj_weight_quantizers.11 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.down_proj_weight_quantizers.12 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.down_proj_weight_quantizers.13 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.down_proj_weight_quantizers.14 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.down_proj_weight_quantizers.15 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.50e-02 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.down_proj_weight_quantizers.16 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.down_proj_weight_quantizers.17 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.69e-01 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.down_proj_weight_quantizers.18 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.down_proj_weight_quantizers.19 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.down_proj_weight_quantizers.20 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.down_proj_weight_quantizers.21 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.down_proj_weight_quantizers.22 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.down_proj_weight_quantizers.23 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.35e-01 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.down_proj_weight_quantizers.24 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.down_proj_weight_quantizers.25 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.45e-02 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.down_proj_weight_quantizers.26 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.54e-02 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.down_proj_weight_quantizers.27 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.down_proj_weight_quantizers.28 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.down_proj_weight_quantizers.29 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.28e-01 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.down_proj_weight_quantizers.30 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.50e-02 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.down_proj_weight_quantizers.31 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.down_proj_weight_quantizers.32 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.down_proj_weight_quantizers.33 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.down_proj_weight_quantizers.34 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.down_proj_weight_quantizers.35 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.down_proj_weight_quantizers.36 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.down_proj_weight_quantizers.37 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.down_proj_weight_quantizers.38 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.down_proj_weight_quantizers.39 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.down_proj_weight_quantizers.40 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.down_proj_weight_quantizers.41 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.down_proj_weight_quantizers.42 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.19e-01 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.down_proj_weight_quantizers.43 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.down_proj_weight_quantizers.44 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.29e-01 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.down_proj_weight_quantizers.45 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.down_proj_weight_quantizers.46 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.down_proj_weight_quantizers.47 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.down_proj_weight_quantizers.48 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.down_proj_weight_quantizers.49 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.down_proj_weight_quantizers.50 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.down_proj_weight_quantizers.51 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.22e-01 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.down_proj_weight_quantizers.52 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.down_proj_weight_quantizers.53 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.down_proj_weight_quantizers.54 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.down_proj_weight_quantizers.55 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.down_proj_weight_quantizers.56 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.down_proj_weight_quantizers.57 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.down_proj_weight_quantizers.58 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.down_proj_weight_quantizers.59 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.down_proj_weight_quantizers.60 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.down_proj_weight_quantizers.61 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.down_proj_weight_quantizers.62 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.down_proj_weight_quantizers.63 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.down_proj_weight_quantizers.64 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.down_proj_weight_quantizers.65 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.down_proj_weight_quantizers.66 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.88e-01 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.down_proj_weight_quantizers.67 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.down_proj_weight_quantizers.68 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.down_proj_weight_quantizers.69 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.down_proj_weight_quantizers.70 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.down_proj_weight_quantizers.71 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.down_proj_weight_quantizers.72 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.down_proj_weight_quantizers.73 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.down_proj_weight_quantizers.74 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.down_proj_weight_quantizers.75 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.down_proj_weight_quantizers.76 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.down_proj_weight_quantizers.77 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.64e-02 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.down_proj_weight_quantizers.78 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.down_proj_weight_quantizers.79 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.down_proj_weight_quantizers.80 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.down_proj_weight_quantizers.81 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.down_proj_weight_quantizers.82 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.down_proj_weight_quantizers.83 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.down_proj_weight_quantizers.84 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.down_proj_weight_quantizers.85 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.down_proj_weight_quantizers.86 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.82e-01 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.down_proj_weight_quantizers.87 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.down_proj_weight_quantizers.88 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.54e-02 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.down_proj_weight_quantizers.89 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.59e-02 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.down_proj_weight_quantizers.90 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.down_proj_weight_quantizers.91 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.down_proj_weight_quantizers.92 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.down_proj_weight_quantizers.93 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.down_proj_weight_quantizers.94 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.down_proj_weight_quantizers.95 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.down_proj_weight_quantizers.96 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.down_proj_weight_quantizers.97 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.down_proj_weight_quantizers.98 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.down_proj_weight_quantizers.99 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.down_proj_weight_quantizers.100 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.down_proj_weight_quantizers.101 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.54e-02 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.down_proj_weight_quantizers.102 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.down_proj_weight_quantizers.103 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=4.30e-01 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.down_proj_weight_quantizers.104 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.down_proj_weight_quantizers.105 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.down_proj_weight_quantizers.106 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.down_proj_weight_quantizers.107 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.14e-01 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.down_proj_weight_quantizers.108 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.down_proj_weight_quantizers.109 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.down_proj_weight_quantizers.110 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.down_proj_weight_quantizers.111 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.down_proj_weight_quantizers.112 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.down_proj_weight_quantizers.113 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.down_proj_weight_quantizers.114 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.down_proj_weight_quantizers.115 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.30e-01 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.down_proj_weight_quantizers.116 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.down_proj_weight_quantizers.117 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.down_proj_weight_quantizers.118 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.down_proj_weight_quantizers.119 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.down_proj_weight_quantizers.120 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.21e-01 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.down_proj_weight_quantizers.121 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.down_proj_weight_quantizers.122 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.30e-01 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.down_proj_weight_quantizers.123 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.down_proj_weight_quantizers.124 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.down_proj_weight_quantizers.125 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.down_proj_weight_quantizers.126 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.down_proj_weight_quantizers.127 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.down_proj_weight_quantizers.128 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.down_proj_weight_quantizers.129 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.down_proj_weight_quantizers.130 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.down_proj_weight_quantizers.131 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.down_proj_weight_quantizers.132 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.down_proj_weight_quantizers.133 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.down_proj_weight_quantizers.134 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.down_proj_weight_quantizers.135 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.down_proj_weight_quantizers.136 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.down_proj_weight_quantizers.137 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.down_proj_weight_quantizers.138 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.down_proj_weight_quantizers.139 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.23e-01 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.down_proj_weight_quantizers.140 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.down_proj_weight_quantizers.141 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.24e-01 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.down_proj_weight_quantizers.142 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.down_proj_weight_quantizers.143 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.55e-01 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.down_proj_weight_quantizers.144 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.down_proj_weight_quantizers.145 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.down_proj_weight_quantizers.146 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.down_proj_weight_quantizers.147 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.down_proj_weight_quantizers.148 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.down_proj_weight_quantizers.149 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.down_proj_weight_quantizers.150 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.down_proj_weight_quantizers.151 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.45e-02 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.down_proj_weight_quantizers.152 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.down_proj_weight_quantizers.153 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.down_proj_weight_quantizers.154 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.down_proj_weight_quantizers.155 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.25e-01 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.down_proj_weight_quantizers.156 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.70e-01 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.down_proj_weight_quantizers.157 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.down_proj_weight_quantizers.158 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.71e-01 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.down_proj_weight_quantizers.159 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.down_proj_weight_quantizers.160 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.64e-02 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.down_proj_weight_quantizers.161 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.down_proj_weight_quantizers.162 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.down_proj_weight_quantizers.163 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.down_proj_weight_quantizers.164 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.down_proj_weight_quantizers.165 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.down_proj_weight_quantizers.166 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.35e-01 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.down_proj_weight_quantizers.167 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.down_proj_weight_quantizers.168 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.down_proj_weight_quantizers.169 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.59e-02 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.down_proj_weight_quantizers.170 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.down_proj_weight_quantizers.171 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.down_proj_weight_quantizers.172 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.down_proj_weight_quantizers.173 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.down_proj_weight_quantizers.174 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.down_proj_weight_quantizers.175 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.down_proj_weight_quantizers.176 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.down_proj_weight_quantizers.177 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.down_proj_weight_quantizers.178 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.down_proj_weight_quantizers.179 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.down_proj_weight_quantizers.180 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.down_proj_weight_quantizers.181 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.14e-01 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.down_proj_weight_quantizers.182 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.34e-01 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.down_proj_weight_quantizers.183 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.down_proj_weight_quantizers.184 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.down_proj_weight_quantizers.185 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.down_proj_weight_quantizers.186 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.down_proj_weight_quantizers.187 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.61e-01 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.down_proj_weight_quantizers.188 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.down_proj_weight_quantizers.189 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.down_proj_weight_quantizers.190 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.47e-01 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.down_proj_weight_quantizers.191 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.down_proj_weight_quantizers.192 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.down_proj_weight_quantizers.193 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.down_proj_weight_quantizers.194 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.down_proj_weight_quantizers.195 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.down_proj_weight_quantizers.196 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.down_proj_weight_quantizers.197 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.38e-01 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.down_proj_weight_quantizers.198 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.down_proj_weight_quantizers.199 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.down_proj_weight_quantizers.200 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.down_proj_weight_quantizers.201 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.down_proj_weight_quantizers.202 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.45e-01 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.down_proj_weight_quantizers.203 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.down_proj_weight_quantizers.204 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.down_proj_weight_quantizers.205 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.down_proj_weight_quantizers.206 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.down_proj_weight_quantizers.207 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.down_proj_weight_quantizers.208 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.50e-02 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.down_proj_weight_quantizers.209 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.down_proj_weight_quantizers.210 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.down_proj_weight_quantizers.211 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.down_proj_weight_quantizers.212 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.down_proj_weight_quantizers.213 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.down_proj_weight_quantizers.214 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.down_proj_weight_quantizers.215 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.down_proj_weight_quantizers.216 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.down_proj_weight_quantizers.217 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.down_proj_weight_quantizers.218 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.down_proj_weight_quantizers.219 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.down_proj_weight_quantizers.220 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.down_proj_weight_quantizers.221 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.down_proj_weight_quantizers.222 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.13e-01 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.down_proj_weight_quantizers.223 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.down_proj_weight_quantizers.224 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.down_proj_weight_quantizers.225 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.down_proj_weight_quantizers.226 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.down_proj_weight_quantizers.227 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.down_proj_weight_quantizers.228 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.down_proj_weight_quantizers.229 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.down_proj_weight_quantizers.230 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.down_proj_weight_quantizers.231 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.down_proj_weight_quantizers.232 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.down_proj_weight_quantizers.233 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.down_proj_weight_quantizers.234 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.45e-01 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.down_proj_weight_quantizers.235 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.down_proj_weight_quantizers.236 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.down_proj_weight_quantizers.237 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.down_proj_weight_quantizers.238 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.down_proj_weight_quantizers.239 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.down_proj_weight_quantizers.240 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.64e-02 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.down_proj_weight_quantizers.241 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.55e-01 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.down_proj_weight_quantizers.242 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.down_proj_weight_quantizers.243 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.down_proj_weight_quantizers.244 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.down_proj_weight_quantizers.245 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.16e-01 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.down_proj_weight_quantizers.246 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.down_proj_weight_quantizers.247 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.50e-02 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.down_proj_weight_quantizers.248 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.down_proj_weight_quantizers.249 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.down_proj_weight_quantizers.250 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.down_proj_weight_quantizers.251 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.down_proj_weight_quantizers.252 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.down_proj_weight_quantizers.253 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.down_proj_weight_quantizers.254 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.down_proj_weight_quantizers.255 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.57.mlp.shared_experts.gate_proj.input_quantizer TensorQuantizer(disabled) +model.layers.57.mlp.shared_experts.gate_proj.output_quantizer TensorQuantizer(disabled) +model.layers.57.mlp.shared_experts.gate_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.57.mlp.shared_experts.up_proj.input_quantizer TensorQuantizer(disabled) +model.layers.57.mlp.shared_experts.up_proj.output_quantizer TensorQuantizer(disabled) +model.layers.57.mlp.shared_experts.up_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.57.mlp.shared_experts.down_proj.input_quantizer TensorQuantizer(disabled) +model.layers.57.mlp.shared_experts.down_proj.output_quantizer TensorQuantizer(disabled) +model.layers.57.mlp.shared_experts.down_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.58.self_attn.q_a_proj.input_quantizer TensorQuantizer(disabled) +model.layers.58.self_attn.q_a_proj.output_quantizer TensorQuantizer(disabled) +model.layers.58.self_attn.q_a_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.58.self_attn.q_b_proj.input_quantizer TensorQuantizer(disabled) +model.layers.58.self_attn.q_b_proj.output_quantizer TensorQuantizer(disabled) +model.layers.58.self_attn.q_b_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.58.self_attn.kv_a_proj_with_mqa.input_quantizer TensorQuantizer(disabled) +model.layers.58.self_attn.kv_a_proj_with_mqa.output_quantizer TensorQuantizer(disabled) +model.layers.58.self_attn.kv_a_proj_with_mqa.weight_quantizer TensorQuantizer(disabled) +model.layers.58.self_attn.kv_b_proj.input_quantizer TensorQuantizer(disabled) +model.layers.58.self_attn.kv_b_proj.output_quantizer TensorQuantizer(disabled) +model.layers.58.self_attn.kv_b_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.58.self_attn.o_proj.input_quantizer TensorQuantizer(disabled) +model.layers.58.self_attn.o_proj.output_quantizer TensorQuantizer(disabled) +model.layers.58.self_attn.o_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.58.self_attn.indexer.wq_b.input_quantizer TensorQuantizer(disabled) +model.layers.58.self_attn.indexer.wq_b.output_quantizer TensorQuantizer(disabled) +model.layers.58.self_attn.indexer.wq_b.weight_quantizer TensorQuantizer(disabled) +model.layers.58.self_attn.indexer.wk.input_quantizer TensorQuantizer(disabled) +model.layers.58.self_attn.indexer.wk.output_quantizer TensorQuantizer(disabled) +model.layers.58.self_attn.indexer.wk.weight_quantizer TensorQuantizer(disabled) +model.layers.58.self_attn.indexer.k_norm.input_quantizer TensorQuantizer(disabled) +model.layers.58.self_attn.indexer.k_norm.output_quantizer TensorQuantizer(disabled) +model.layers.58.self_attn.indexer.weights_proj.input_quantizer TensorQuantizer(disabled) +model.layers.58.self_attn.indexer.weights_proj.output_quantizer TensorQuantizer(disabled) +model.layers.58.self_attn.indexer.weights_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.58.self_attn.q_bmm_quantizer TensorQuantizer(disabled) +model.layers.58.self_attn.k_bmm_quantizer TensorQuantizer((4, 3) bit fake per-tensor amax=4.48e+02(const) calibrator=MaxCalibrator quant) +model.layers.58.self_attn.v_bmm_quantizer TensorQuantizer((4, 3) bit fake per-tensor amax=4.48e+02(const) calibrator=MaxCalibrator quant) +model.layers.58.self_attn.p_bmm_quantizer TensorQuantizer(disabled) +model.layers.58.mlp.experts.gate_up_proj_input_quantizer TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.92e+00 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.gate_up_proj_weight_quantizers.0 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.gate_up_proj_weight_quantizers.1 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.gate_up_proj_weight_quantizers.2 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.23e-01 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.gate_up_proj_weight_quantizers.3 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.gate_up_proj_weight_quantizers.4 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.gate_up_proj_weight_quantizers.5 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.13e-01 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.gate_up_proj_weight_quantizers.6 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.gate_up_proj_weight_quantizers.7 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.gate_up_proj_weight_quantizers.8 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.gate_up_proj_weight_quantizers.9 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.24e-01 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.gate_up_proj_weight_quantizers.10 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.65e-01 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.gate_up_proj_weight_quantizers.11 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.gate_up_proj_weight_quantizers.12 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.gate_up_proj_weight_quantizers.13 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.gate_up_proj_weight_quantizers.14 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.gate_up_proj_weight_quantizers.15 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.gate_up_proj_weight_quantizers.16 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.gate_up_proj_weight_quantizers.17 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.gate_up_proj_weight_quantizers.18 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.gate_up_proj_weight_quantizers.19 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.gate_up_proj_weight_quantizers.20 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.gate_up_proj_weight_quantizers.21 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.43e-01 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.gate_up_proj_weight_quantizers.22 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.gate_up_proj_weight_quantizers.23 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.gate_up_proj_weight_quantizers.24 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.gate_up_proj_weight_quantizers.25 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.gate_up_proj_weight_quantizers.26 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.36e-01 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.gate_up_proj_weight_quantizers.27 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.gate_up_proj_weight_quantizers.28 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.17e-01 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.gate_up_proj_weight_quantizers.29 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.gate_up_proj_weight_quantizers.30 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.gate_up_proj_weight_quantizers.31 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.gate_up_proj_weight_quantizers.32 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.gate_up_proj_weight_quantizers.33 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.gate_up_proj_weight_quantizers.34 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.18e-01 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.gate_up_proj_weight_quantizers.35 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.07e-01 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.gate_up_proj_weight_quantizers.36 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.16e-01 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.gate_up_proj_weight_quantizers.37 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.gate_up_proj_weight_quantizers.38 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.gate_up_proj_weight_quantizers.39 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.gate_up_proj_weight_quantizers.40 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.gate_up_proj_weight_quantizers.41 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.gate_up_proj_weight_quantizers.42 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.13e-01 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.gate_up_proj_weight_quantizers.43 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.gate_up_proj_weight_quantizers.44 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.gate_up_proj_weight_quantizers.45 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.gate_up_proj_weight_quantizers.46 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.14e-01 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.gate_up_proj_weight_quantizers.47 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.gate_up_proj_weight_quantizers.48 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.gate_up_proj_weight_quantizers.49 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.gate_up_proj_weight_quantizers.50 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.gate_up_proj_weight_quantizers.51 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.19e-01 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.gate_up_proj_weight_quantizers.52 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.gate_up_proj_weight_quantizers.53 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.gate_up_proj_weight_quantizers.54 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.gate_up_proj_weight_quantizers.55 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.gate_up_proj_weight_quantizers.56 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.gate_up_proj_weight_quantizers.57 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.gate_up_proj_weight_quantizers.58 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.gate_up_proj_weight_quantizers.59 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.gate_up_proj_weight_quantizers.60 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.gate_up_proj_weight_quantizers.61 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.gate_up_proj_weight_quantizers.62 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.gate_up_proj_weight_quantizers.63 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.25e-01 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.gate_up_proj_weight_quantizers.64 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.gate_up_proj_weight_quantizers.65 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.gate_up_proj_weight_quantizers.66 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.gate_up_proj_weight_quantizers.67 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.gate_up_proj_weight_quantizers.68 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.gate_up_proj_weight_quantizers.69 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.gate_up_proj_weight_quantizers.70 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.gate_up_proj_weight_quantizers.71 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.28e-01 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.gate_up_proj_weight_quantizers.72 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.43e-01 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.gate_up_proj_weight_quantizers.73 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.74e-01 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.gate_up_proj_weight_quantizers.74 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.gate_up_proj_weight_quantizers.75 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.18e-01 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.gate_up_proj_weight_quantizers.76 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.gate_up_proj_weight_quantizers.77 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.gate_up_proj_weight_quantizers.78 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.gate_up_proj_weight_quantizers.79 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.gate_up_proj_weight_quantizers.80 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.gate_up_proj_weight_quantizers.81 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.gate_up_proj_weight_quantizers.82 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.gate_up_proj_weight_quantizers.83 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.gate_up_proj_weight_quantizers.84 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.gate_up_proj_weight_quantizers.85 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.99e-01 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.gate_up_proj_weight_quantizers.86 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.gate_up_proj_weight_quantizers.87 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.gate_up_proj_weight_quantizers.88 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.gate_up_proj_weight_quantizers.89 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.gate_up_proj_weight_quantizers.90 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.gate_up_proj_weight_quantizers.91 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.23e-01 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.gate_up_proj_weight_quantizers.92 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.17e-01 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.gate_up_proj_weight_quantizers.93 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.gate_up_proj_weight_quantizers.94 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.gate_up_proj_weight_quantizers.95 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.gate_up_proj_weight_quantizers.96 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.20e-01 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.gate_up_proj_weight_quantizers.97 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.gate_up_proj_weight_quantizers.98 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.gate_up_proj_weight_quantizers.99 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.19e-01 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.gate_up_proj_weight_quantizers.100 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.14e-01 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.gate_up_proj_weight_quantizers.101 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.gate_up_proj_weight_quantizers.102 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.gate_up_proj_weight_quantizers.103 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.gate_up_proj_weight_quantizers.104 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.gate_up_proj_weight_quantizers.105 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.gate_up_proj_weight_quantizers.106 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.gate_up_proj_weight_quantizers.107 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.gate_up_proj_weight_quantizers.108 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.gate_up_proj_weight_quantizers.109 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.gate_up_proj_weight_quantizers.110 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.gate_up_proj_weight_quantizers.111 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.gate_up_proj_weight_quantizers.112 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.gate_up_proj_weight_quantizers.113 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.14e-01 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.gate_up_proj_weight_quantizers.114 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.gate_up_proj_weight_quantizers.115 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.gate_up_proj_weight_quantizers.116 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.gate_up_proj_weight_quantizers.117 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.16e-01 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.gate_up_proj_weight_quantizers.118 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.gate_up_proj_weight_quantizers.119 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.gate_up_proj_weight_quantizers.120 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.85e-01 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.gate_up_proj_weight_quantizers.121 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.92e-01 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.gate_up_proj_weight_quantizers.122 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.gate_up_proj_weight_quantizers.123 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.20e-01 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.gate_up_proj_weight_quantizers.124 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.gate_up_proj_weight_quantizers.125 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.gate_up_proj_weight_quantizers.126 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.gate_up_proj_weight_quantizers.127 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.gate_up_proj_weight_quantizers.128 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.gate_up_proj_weight_quantizers.129 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.gate_up_proj_weight_quantizers.130 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.gate_up_proj_weight_quantizers.131 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.gate_up_proj_weight_quantizers.132 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.gate_up_proj_weight_quantizers.133 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.gate_up_proj_weight_quantizers.134 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.gate_up_proj_weight_quantizers.135 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.gate_up_proj_weight_quantizers.136 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.gate_up_proj_weight_quantizers.137 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.gate_up_proj_weight_quantizers.138 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.gate_up_proj_weight_quantizers.139 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.gate_up_proj_weight_quantizers.140 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.gate_up_proj_weight_quantizers.141 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.gate_up_proj_weight_quantizers.142 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.37e-01 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.gate_up_proj_weight_quantizers.143 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.gate_up_proj_weight_quantizers.144 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.gate_up_proj_weight_quantizers.145 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.gate_up_proj_weight_quantizers.146 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.gate_up_proj_weight_quantizers.147 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.gate_up_proj_weight_quantizers.148 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.gate_up_proj_weight_quantizers.149 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.14e-01 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.gate_up_proj_weight_quantizers.150 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.gate_up_proj_weight_quantizers.151 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.gate_up_proj_weight_quantizers.152 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.gate_up_proj_weight_quantizers.153 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.gate_up_proj_weight_quantizers.154 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.gate_up_proj_weight_quantizers.155 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.gate_up_proj_weight_quantizers.156 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.20e-01 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.gate_up_proj_weight_quantizers.157 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.18e-01 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.gate_up_proj_weight_quantizers.158 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.gate_up_proj_weight_quantizers.159 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.gate_up_proj_weight_quantizers.160 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.97e-01 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.gate_up_proj_weight_quantizers.161 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.gate_up_proj_weight_quantizers.162 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.gate_up_proj_weight_quantizers.163 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.gate_up_proj_weight_quantizers.164 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.gate_up_proj_weight_quantizers.165 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.gate_up_proj_weight_quantizers.166 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.gate_up_proj_weight_quantizers.167 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.gate_up_proj_weight_quantizers.168 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.gate_up_proj_weight_quantizers.169 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.gate_up_proj_weight_quantizers.170 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.20e-01 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.gate_up_proj_weight_quantizers.171 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.gate_up_proj_weight_quantizers.172 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.gate_up_proj_weight_quantizers.173 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.gate_up_proj_weight_quantizers.174 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.gate_up_proj_weight_quantizers.175 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.19e-01 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.gate_up_proj_weight_quantizers.176 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.gate_up_proj_weight_quantizers.177 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.16e-01 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.gate_up_proj_weight_quantizers.178 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.82e-01 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.gate_up_proj_weight_quantizers.179 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.gate_up_proj_weight_quantizers.180 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.16e-01 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.gate_up_proj_weight_quantizers.181 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.gate_up_proj_weight_quantizers.182 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.gate_up_proj_weight_quantizers.183 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.gate_up_proj_weight_quantizers.184 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.gate_up_proj_weight_quantizers.185 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.gate_up_proj_weight_quantizers.186 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.gate_up_proj_weight_quantizers.187 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.gate_up_proj_weight_quantizers.188 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.gate_up_proj_weight_quantizers.189 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.gate_up_proj_weight_quantizers.190 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.gate_up_proj_weight_quantizers.191 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.60e-01 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.gate_up_proj_weight_quantizers.192 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.36e-01 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.gate_up_proj_weight_quantizers.193 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.03e-01 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.gate_up_proj_weight_quantizers.194 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.gate_up_proj_weight_quantizers.195 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.27e-01 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.gate_up_proj_weight_quantizers.196 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.gate_up_proj_weight_quantizers.197 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.gate_up_proj_weight_quantizers.198 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.gate_up_proj_weight_quantizers.199 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.gate_up_proj_weight_quantizers.200 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.gate_up_proj_weight_quantizers.201 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.gate_up_proj_weight_quantizers.202 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.gate_up_proj_weight_quantizers.203 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.gate_up_proj_weight_quantizers.204 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.gate_up_proj_weight_quantizers.205 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.gate_up_proj_weight_quantizers.206 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.19e-01 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.gate_up_proj_weight_quantizers.207 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.gate_up_proj_weight_quantizers.208 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.gate_up_proj_weight_quantizers.209 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.80e-01 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.gate_up_proj_weight_quantizers.210 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.gate_up_proj_weight_quantizers.211 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.gate_up_proj_weight_quantizers.212 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.gate_up_proj_weight_quantizers.213 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.gate_up_proj_weight_quantizers.214 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.18e-01 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.gate_up_proj_weight_quantizers.215 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.gate_up_proj_weight_quantizers.216 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.gate_up_proj_weight_quantizers.217 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.gate_up_proj_weight_quantizers.218 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.gate_up_proj_weight_quantizers.219 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.35e-01 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.gate_up_proj_weight_quantizers.220 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.gate_up_proj_weight_quantizers.221 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.gate_up_proj_weight_quantizers.222 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.gate_up_proj_weight_quantizers.223 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.13e-01 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.gate_up_proj_weight_quantizers.224 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.gate_up_proj_weight_quantizers.225 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.gate_up_proj_weight_quantizers.226 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.gate_up_proj_weight_quantizers.227 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.gate_up_proj_weight_quantizers.228 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.gate_up_proj_weight_quantizers.229 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.gate_up_proj_weight_quantizers.230 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.gate_up_proj_weight_quantizers.231 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.gate_up_proj_weight_quantizers.232 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.gate_up_proj_weight_quantizers.233 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.33e-01 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.gate_up_proj_weight_quantizers.234 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.gate_up_proj_weight_quantizers.235 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.gate_up_proj_weight_quantizers.236 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.gate_up_proj_weight_quantizers.237 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.gate_up_proj_weight_quantizers.238 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.18e-01 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.gate_up_proj_weight_quantizers.239 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.gate_up_proj_weight_quantizers.240 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.79e-01 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.gate_up_proj_weight_quantizers.241 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.gate_up_proj_weight_quantizers.242 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.gate_up_proj_weight_quantizers.243 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.gate_up_proj_weight_quantizers.244 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.gate_up_proj_weight_quantizers.245 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.19e-01 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.gate_up_proj_weight_quantizers.246 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.gate_up_proj_weight_quantizers.247 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.26e-01 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.gate_up_proj_weight_quantizers.248 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.gate_up_proj_weight_quantizers.249 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.97e-01 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.gate_up_proj_weight_quantizers.250 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.gate_up_proj_weight_quantizers.251 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.23e-01 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.gate_up_proj_weight_quantizers.252 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.gate_up_proj_weight_quantizers.253 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.08e-01 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.gate_up_proj_weight_quantizers.254 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.gate_up_proj_weight_quantizers.255 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.down_proj_input_quantizer TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=5.25e+01 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.down_proj_weight_quantizers.0 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.down_proj_weight_quantizers.1 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.down_proj_weight_quantizers.2 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.down_proj_weight_quantizers.3 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.down_proj_weight_quantizers.4 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.down_proj_weight_quantizers.5 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.down_proj_weight_quantizers.6 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.down_proj_weight_quantizers.7 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.down_proj_weight_quantizers.8 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.down_proj_weight_quantizers.9 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.down_proj_weight_quantizers.10 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.14e-01 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.down_proj_weight_quantizers.11 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.down_proj_weight_quantizers.12 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.down_proj_weight_quantizers.13 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.down_proj_weight_quantizers.14 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.down_proj_weight_quantizers.15 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.down_proj_weight_quantizers.16 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.down_proj_weight_quantizers.17 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.down_proj_weight_quantizers.18 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.down_proj_weight_quantizers.19 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.down_proj_weight_quantizers.20 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.down_proj_weight_quantizers.21 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.69e-01 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.down_proj_weight_quantizers.22 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.down_proj_weight_quantizers.23 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.down_proj_weight_quantizers.24 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.down_proj_weight_quantizers.25 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.down_proj_weight_quantizers.26 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.down_proj_weight_quantizers.27 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.down_proj_weight_quantizers.28 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.down_proj_weight_quantizers.29 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.down_proj_weight_quantizers.30 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.down_proj_weight_quantizers.31 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.down_proj_weight_quantizers.32 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.down_proj_weight_quantizers.33 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.down_proj_weight_quantizers.34 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.81e-01 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.down_proj_weight_quantizers.35 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.down_proj_weight_quantizers.36 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.down_proj_weight_quantizers.37 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.25e-02 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.down_proj_weight_quantizers.38 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.down_proj_weight_quantizers.39 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.down_proj_weight_quantizers.40 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.down_proj_weight_quantizers.41 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.down_proj_weight_quantizers.42 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.down_proj_weight_quantizers.43 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.down_proj_weight_quantizers.44 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.down_proj_weight_quantizers.45 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.down_proj_weight_quantizers.46 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.down_proj_weight_quantizers.47 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.down_proj_weight_quantizers.48 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.down_proj_weight_quantizers.49 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.24e-01 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.down_proj_weight_quantizers.50 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.down_proj_weight_quantizers.51 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.down_proj_weight_quantizers.52 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.down_proj_weight_quantizers.53 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.54e-02 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.down_proj_weight_quantizers.54 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.down_proj_weight_quantizers.55 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.down_proj_weight_quantizers.56 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.down_proj_weight_quantizers.57 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.down_proj_weight_quantizers.58 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.67e-01 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.down_proj_weight_quantizers.59 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.down_proj_weight_quantizers.60 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.down_proj_weight_quantizers.61 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.down_proj_weight_quantizers.62 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.down_proj_weight_quantizers.63 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.49e-01 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.down_proj_weight_quantizers.64 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.down_proj_weight_quantizers.65 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.down_proj_weight_quantizers.66 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.down_proj_weight_quantizers.67 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.37e-01 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.down_proj_weight_quantizers.68 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.down_proj_weight_quantizers.69 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.down_proj_weight_quantizers.70 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.29e-01 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.down_proj_weight_quantizers.71 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.down_proj_weight_quantizers.72 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.29e-01 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.down_proj_weight_quantizers.73 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.down_proj_weight_quantizers.74 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.down_proj_weight_quantizers.75 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.87e-01 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.down_proj_weight_quantizers.76 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.down_proj_weight_quantizers.77 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.down_proj_weight_quantizers.78 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.55e-01 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.down_proj_weight_quantizers.79 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.down_proj_weight_quantizers.80 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.down_proj_weight_quantizers.81 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.down_proj_weight_quantizers.82 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.down_proj_weight_quantizers.83 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.down_proj_weight_quantizers.84 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.down_proj_weight_quantizers.85 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.34e-01 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.down_proj_weight_quantizers.86 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.down_proj_weight_quantizers.87 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.down_proj_weight_quantizers.88 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.down_proj_weight_quantizers.89 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.32e-01 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.down_proj_weight_quantizers.90 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.down_proj_weight_quantizers.91 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.down_proj_weight_quantizers.92 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.down_proj_weight_quantizers.93 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.down_proj_weight_quantizers.94 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.down_proj_weight_quantizers.95 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.down_proj_weight_quantizers.96 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.down_proj_weight_quantizers.97 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.down_proj_weight_quantizers.98 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.down_proj_weight_quantizers.99 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.down_proj_weight_quantizers.100 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.down_proj_weight_quantizers.101 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.down_proj_weight_quantizers.102 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.down_proj_weight_quantizers.103 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.down_proj_weight_quantizers.104 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.64e-02 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.down_proj_weight_quantizers.105 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.down_proj_weight_quantizers.106 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.down_proj_weight_quantizers.107 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.down_proj_weight_quantizers.108 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.down_proj_weight_quantizers.109 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.down_proj_weight_quantizers.110 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.down_proj_weight_quantizers.111 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.25e-01 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.down_proj_weight_quantizers.112 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.down_proj_weight_quantizers.113 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.59e-01 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.down_proj_weight_quantizers.114 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.down_proj_weight_quantizers.115 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.64e-02 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.down_proj_weight_quantizers.116 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.down_proj_weight_quantizers.117 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.down_proj_weight_quantizers.118 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.down_proj_weight_quantizers.119 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.down_proj_weight_quantizers.120 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.81e-01 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.down_proj_weight_quantizers.121 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.down_proj_weight_quantizers.122 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.down_proj_weight_quantizers.123 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.down_proj_weight_quantizers.124 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.down_proj_weight_quantizers.125 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.33e-01 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.down_proj_weight_quantizers.126 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.down_proj_weight_quantizers.127 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.down_proj_weight_quantizers.128 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.down_proj_weight_quantizers.129 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.down_proj_weight_quantizers.130 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.64e-02 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.down_proj_weight_quantizers.131 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.89e-01 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.down_proj_weight_quantizers.132 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.down_proj_weight_quantizers.133 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.down_proj_weight_quantizers.134 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.down_proj_weight_quantizers.135 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.down_proj_weight_quantizers.136 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.down_proj_weight_quantizers.137 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.down_proj_weight_quantizers.138 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.down_proj_weight_quantizers.139 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.down_proj_weight_quantizers.140 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.down_proj_weight_quantizers.141 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.down_proj_weight_quantizers.142 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.down_proj_weight_quantizers.143 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.down_proj_weight_quantizers.144 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.down_proj_weight_quantizers.145 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.down_proj_weight_quantizers.146 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.down_proj_weight_quantizers.147 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.down_proj_weight_quantizers.148 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.down_proj_weight_quantizers.149 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.down_proj_weight_quantizers.150 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.down_proj_weight_quantizers.151 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.down_proj_weight_quantizers.152 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.down_proj_weight_quantizers.153 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.down_proj_weight_quantizers.154 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.down_proj_weight_quantizers.155 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.39e-01 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.down_proj_weight_quantizers.156 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.down_proj_weight_quantizers.157 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.down_proj_weight_quantizers.158 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.down_proj_weight_quantizers.159 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.down_proj_weight_quantizers.160 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.down_proj_weight_quantizers.161 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.down_proj_weight_quantizers.162 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.down_proj_weight_quantizers.163 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.45e-01 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.down_proj_weight_quantizers.164 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.down_proj_weight_quantizers.165 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.down_proj_weight_quantizers.166 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.down_proj_weight_quantizers.167 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.down_proj_weight_quantizers.168 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.down_proj_weight_quantizers.169 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.down_proj_weight_quantizers.170 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.32e-01 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.down_proj_weight_quantizers.171 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.down_proj_weight_quantizers.172 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.43e-01 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.down_proj_weight_quantizers.173 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.down_proj_weight_quantizers.174 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.down_proj_weight_quantizers.175 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.down_proj_weight_quantizers.176 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.39e-01 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.down_proj_weight_quantizers.177 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.down_proj_weight_quantizers.178 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.50e-01 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.down_proj_weight_quantizers.179 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.down_proj_weight_quantizers.180 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.down_proj_weight_quantizers.181 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.down_proj_weight_quantizers.182 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.down_proj_weight_quantizers.183 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.29e-01 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.down_proj_weight_quantizers.184 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.down_proj_weight_quantizers.185 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.down_proj_weight_quantizers.186 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.down_proj_weight_quantizers.187 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.down_proj_weight_quantizers.188 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.down_proj_weight_quantizers.189 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.down_proj_weight_quantizers.190 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.down_proj_weight_quantizers.191 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.42e-01 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.down_proj_weight_quantizers.192 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.91e-01 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.down_proj_weight_quantizers.193 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.down_proj_weight_quantizers.194 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.14e-01 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.down_proj_weight_quantizers.195 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.down_proj_weight_quantizers.196 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.down_proj_weight_quantizers.197 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.down_proj_weight_quantizers.198 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.64e-02 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.down_proj_weight_quantizers.199 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.down_proj_weight_quantizers.200 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.down_proj_weight_quantizers.201 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.down_proj_weight_quantizers.202 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.down_proj_weight_quantizers.203 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.15e-01 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.down_proj_weight_quantizers.204 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.down_proj_weight_quantizers.205 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.down_proj_weight_quantizers.206 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.down_proj_weight_quantizers.207 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.down_proj_weight_quantizers.208 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.down_proj_weight_quantizers.209 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.down_proj_weight_quantizers.210 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.down_proj_weight_quantizers.211 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.19e-01 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.down_proj_weight_quantizers.212 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.down_proj_weight_quantizers.213 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.down_proj_weight_quantizers.214 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.down_proj_weight_quantizers.215 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.down_proj_weight_quantizers.216 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.down_proj_weight_quantizers.217 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.down_proj_weight_quantizers.218 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.down_proj_weight_quantizers.219 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.07e-01 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.down_proj_weight_quantizers.220 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.down_proj_weight_quantizers.221 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.down_proj_weight_quantizers.222 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.down_proj_weight_quantizers.223 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.down_proj_weight_quantizers.224 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.32e-01 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.down_proj_weight_quantizers.225 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.down_proj_weight_quantizers.226 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.down_proj_weight_quantizers.227 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.down_proj_weight_quantizers.228 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.down_proj_weight_quantizers.229 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.64e-02 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.down_proj_weight_quantizers.230 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.down_proj_weight_quantizers.231 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.down_proj_weight_quantizers.232 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.down_proj_weight_quantizers.233 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.down_proj_weight_quantizers.234 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.down_proj_weight_quantizers.235 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.down_proj_weight_quantizers.236 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.down_proj_weight_quantizers.237 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.down_proj_weight_quantizers.238 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.down_proj_weight_quantizers.239 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.down_proj_weight_quantizers.240 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.down_proj_weight_quantizers.241 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.down_proj_weight_quantizers.242 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.down_proj_weight_quantizers.243 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.down_proj_weight_quantizers.244 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.down_proj_weight_quantizers.245 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.down_proj_weight_quantizers.246 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.59e-02 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.down_proj_weight_quantizers.247 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.down_proj_weight_quantizers.248 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.down_proj_weight_quantizers.249 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.25e-01 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.down_proj_weight_quantizers.250 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.down_proj_weight_quantizers.251 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.28e-01 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.down_proj_weight_quantizers.252 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.59e-02 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.down_proj_weight_quantizers.253 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.00e-01 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.down_proj_weight_quantizers.254 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.down_proj_weight_quantizers.255 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.58.mlp.shared_experts.gate_proj.input_quantizer TensorQuantizer(disabled) +model.layers.58.mlp.shared_experts.gate_proj.output_quantizer TensorQuantizer(disabled) +model.layers.58.mlp.shared_experts.gate_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.58.mlp.shared_experts.up_proj.input_quantizer TensorQuantizer(disabled) +model.layers.58.mlp.shared_experts.up_proj.output_quantizer TensorQuantizer(disabled) +model.layers.58.mlp.shared_experts.up_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.58.mlp.shared_experts.down_proj.input_quantizer TensorQuantizer(disabled) +model.layers.58.mlp.shared_experts.down_proj.output_quantizer TensorQuantizer(disabled) +model.layers.58.mlp.shared_experts.down_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.59.self_attn.q_a_proj.input_quantizer TensorQuantizer(disabled) +model.layers.59.self_attn.q_a_proj.output_quantizer TensorQuantizer(disabled) +model.layers.59.self_attn.q_a_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.59.self_attn.q_b_proj.input_quantizer TensorQuantizer(disabled) +model.layers.59.self_attn.q_b_proj.output_quantizer TensorQuantizer(disabled) +model.layers.59.self_attn.q_b_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.59.self_attn.kv_a_proj_with_mqa.input_quantizer TensorQuantizer(disabled) +model.layers.59.self_attn.kv_a_proj_with_mqa.output_quantizer TensorQuantizer(disabled) +model.layers.59.self_attn.kv_a_proj_with_mqa.weight_quantizer TensorQuantizer(disabled) +model.layers.59.self_attn.kv_b_proj.input_quantizer TensorQuantizer(disabled) +model.layers.59.self_attn.kv_b_proj.output_quantizer TensorQuantizer(disabled) +model.layers.59.self_attn.kv_b_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.59.self_attn.o_proj.input_quantizer TensorQuantizer(disabled) +model.layers.59.self_attn.o_proj.output_quantizer TensorQuantizer(disabled) +model.layers.59.self_attn.o_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.59.self_attn.q_bmm_quantizer TensorQuantizer(disabled) +model.layers.59.self_attn.k_bmm_quantizer TensorQuantizer((4, 3) bit fake per-tensor amax=4.48e+02(const) calibrator=MaxCalibrator quant) +model.layers.59.self_attn.v_bmm_quantizer TensorQuantizer((4, 3) bit fake per-tensor amax=4.48e+02(const) calibrator=MaxCalibrator quant) +model.layers.59.self_attn.p_bmm_quantizer TensorQuantizer(disabled) +model.layers.59.mlp.experts.gate_up_proj_input_quantizer TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.64e+00 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.gate_up_proj_weight_quantizers.0 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.gate_up_proj_weight_quantizers.1 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.gate_up_proj_weight_quantizers.2 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.gate_up_proj_weight_quantizers.3 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.14e-01 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.gate_up_proj_weight_quantizers.4 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.gate_up_proj_weight_quantizers.5 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.26e-01 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.gate_up_proj_weight_quantizers.6 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.gate_up_proj_weight_quantizers.7 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.14e-01 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.gate_up_proj_weight_quantizers.8 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.23e-01 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.gate_up_proj_weight_quantizers.9 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.gate_up_proj_weight_quantizers.10 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.gate_up_proj_weight_quantizers.11 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.gate_up_proj_weight_quantizers.12 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.gate_up_proj_weight_quantizers.13 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.25e-01 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.gate_up_proj_weight_quantizers.14 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.gate_up_proj_weight_quantizers.15 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.gate_up_proj_weight_quantizers.16 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.gate_up_proj_weight_quantizers.17 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.gate_up_proj_weight_quantizers.18 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.46e-01 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.gate_up_proj_weight_quantizers.19 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.gate_up_proj_weight_quantizers.20 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.gate_up_proj_weight_quantizers.21 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.gate_up_proj_weight_quantizers.22 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.gate_up_proj_weight_quantizers.23 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.gate_up_proj_weight_quantizers.24 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.gate_up_proj_weight_quantizers.25 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.gate_up_proj_weight_quantizers.26 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.gate_up_proj_weight_quantizers.27 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.24e-01 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.gate_up_proj_weight_quantizers.28 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.70e-01 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.gate_up_proj_weight_quantizers.29 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.gate_up_proj_weight_quantizers.30 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.66e-01 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.gate_up_proj_weight_quantizers.31 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.gate_up_proj_weight_quantizers.32 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.gate_up_proj_weight_quantizers.33 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.gate_up_proj_weight_quantizers.34 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.gate_up_proj_weight_quantizers.35 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.gate_up_proj_weight_quantizers.36 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.54e-01 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.gate_up_proj_weight_quantizers.37 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.gate_up_proj_weight_quantizers.38 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.gate_up_proj_weight_quantizers.39 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=4.10e-01 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.gate_up_proj_weight_quantizers.40 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.gate_up_proj_weight_quantizers.41 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.gate_up_proj_weight_quantizers.42 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.42e-01 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.gate_up_proj_weight_quantizers.43 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.42e-01 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.gate_up_proj_weight_quantizers.44 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.gate_up_proj_weight_quantizers.45 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.gate_up_proj_weight_quantizers.46 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.gate_up_proj_weight_quantizers.47 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.gate_up_proj_weight_quantizers.48 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.gate_up_proj_weight_quantizers.49 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.66e-01 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.gate_up_proj_weight_quantizers.50 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.18e-01 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.gate_up_proj_weight_quantizers.51 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.gate_up_proj_weight_quantizers.52 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.gate_up_proj_weight_quantizers.53 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.31e-01 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.gate_up_proj_weight_quantizers.54 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.gate_up_proj_weight_quantizers.55 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.gate_up_proj_weight_quantizers.56 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.13e-01 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.gate_up_proj_weight_quantizers.57 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.gate_up_proj_weight_quantizers.58 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.gate_up_proj_weight_quantizers.59 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.gate_up_proj_weight_quantizers.60 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.gate_up_proj_weight_quantizers.61 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.gate_up_proj_weight_quantizers.62 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.gate_up_proj_weight_quantizers.63 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.gate_up_proj_weight_quantizers.64 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.gate_up_proj_weight_quantizers.65 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.gate_up_proj_weight_quantizers.66 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.gate_up_proj_weight_quantizers.67 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.gate_up_proj_weight_quantizers.68 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.80e-01 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.gate_up_proj_weight_quantizers.69 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.gate_up_proj_weight_quantizers.70 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.gate_up_proj_weight_quantizers.71 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.gate_up_proj_weight_quantizers.72 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.13e-01 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.gate_up_proj_weight_quantizers.73 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.gate_up_proj_weight_quantizers.74 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.gate_up_proj_weight_quantizers.75 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.28e-01 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.gate_up_proj_weight_quantizers.76 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.25e-01 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.gate_up_proj_weight_quantizers.77 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.gate_up_proj_weight_quantizers.78 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.88e-01 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.gate_up_proj_weight_quantizers.79 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.gate_up_proj_weight_quantizers.80 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.gate_up_proj_weight_quantizers.81 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.gate_up_proj_weight_quantizers.82 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.gate_up_proj_weight_quantizers.83 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.gate_up_proj_weight_quantizers.84 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.gate_up_proj_weight_quantizers.85 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.36e-01 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.gate_up_proj_weight_quantizers.86 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.47e-01 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.gate_up_proj_weight_quantizers.87 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.gate_up_proj_weight_quantizers.88 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.gate_up_proj_weight_quantizers.89 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.04e-01 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.gate_up_proj_weight_quantizers.90 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.gate_up_proj_weight_quantizers.91 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.gate_up_proj_weight_quantizers.92 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.gate_up_proj_weight_quantizers.93 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.gate_up_proj_weight_quantizers.94 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.54e-01 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.gate_up_proj_weight_quantizers.95 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.gate_up_proj_weight_quantizers.96 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.gate_up_proj_weight_quantizers.97 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.gate_up_proj_weight_quantizers.98 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.gate_up_proj_weight_quantizers.99 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.gate_up_proj_weight_quantizers.100 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.gate_up_proj_weight_quantizers.101 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.17e-01 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.gate_up_proj_weight_quantizers.102 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.77e-01 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.gate_up_proj_weight_quantizers.103 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.gate_up_proj_weight_quantizers.104 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.gate_up_proj_weight_quantizers.105 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.gate_up_proj_weight_quantizers.106 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.gate_up_proj_weight_quantizers.107 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.gate_up_proj_weight_quantizers.108 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.gate_up_proj_weight_quantizers.109 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.gate_up_proj_weight_quantizers.110 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.gate_up_proj_weight_quantizers.111 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.gate_up_proj_weight_quantizers.112 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.gate_up_proj_weight_quantizers.113 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.gate_up_proj_weight_quantizers.114 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.55e-01 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.gate_up_proj_weight_quantizers.115 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.gate_up_proj_weight_quantizers.116 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.gate_up_proj_weight_quantizers.117 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.gate_up_proj_weight_quantizers.118 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.48e-01 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.gate_up_proj_weight_quantizers.119 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.58e-01 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.gate_up_proj_weight_quantizers.120 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.gate_up_proj_weight_quantizers.121 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.gate_up_proj_weight_quantizers.122 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.gate_up_proj_weight_quantizers.123 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.12e-01 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.gate_up_proj_weight_quantizers.124 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.gate_up_proj_weight_quantizers.125 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.gate_up_proj_weight_quantizers.126 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.25e-01 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.gate_up_proj_weight_quantizers.127 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.44e-01 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.gate_up_proj_weight_quantizers.128 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.gate_up_proj_weight_quantizers.129 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.gate_up_proj_weight_quantizers.130 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.gate_up_proj_weight_quantizers.131 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.gate_up_proj_weight_quantizers.132 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.52e-01 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.gate_up_proj_weight_quantizers.133 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.gate_up_proj_weight_quantizers.134 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.gate_up_proj_weight_quantizers.135 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.gate_up_proj_weight_quantizers.136 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.gate_up_proj_weight_quantizers.137 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.gate_up_proj_weight_quantizers.138 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.gate_up_proj_weight_quantizers.139 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.gate_up_proj_weight_quantizers.140 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.gate_up_proj_weight_quantizers.141 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.25e-01 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.gate_up_proj_weight_quantizers.142 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.gate_up_proj_weight_quantizers.143 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.gate_up_proj_weight_quantizers.144 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.gate_up_proj_weight_quantizers.145 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.17e-01 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.gate_up_proj_weight_quantizers.146 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.gate_up_proj_weight_quantizers.147 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.gate_up_proj_weight_quantizers.148 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.gate_up_proj_weight_quantizers.149 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.gate_up_proj_weight_quantizers.150 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.17e-01 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.gate_up_proj_weight_quantizers.151 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.gate_up_proj_weight_quantizers.152 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.gate_up_proj_weight_quantizers.153 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.gate_up_proj_weight_quantizers.154 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.gate_up_proj_weight_quantizers.155 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.gate_up_proj_weight_quantizers.156 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.gate_up_proj_weight_quantizers.157 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.gate_up_proj_weight_quantizers.158 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.22e-01 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.gate_up_proj_weight_quantizers.159 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.gate_up_proj_weight_quantizers.160 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.gate_up_proj_weight_quantizers.161 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.88e-01 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.gate_up_proj_weight_quantizers.162 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.42e-01 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.gate_up_proj_weight_quantizers.163 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.gate_up_proj_weight_quantizers.164 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.55e-01 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.gate_up_proj_weight_quantizers.165 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.gate_up_proj_weight_quantizers.166 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.49e-01 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.gate_up_proj_weight_quantizers.167 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.gate_up_proj_weight_quantizers.168 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.gate_up_proj_weight_quantizers.169 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.13e-01 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.gate_up_proj_weight_quantizers.170 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.gate_up_proj_weight_quantizers.171 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.gate_up_proj_weight_quantizers.172 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.gate_up_proj_weight_quantizers.173 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.gate_up_proj_weight_quantizers.174 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.gate_up_proj_weight_quantizers.175 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.gate_up_proj_weight_quantizers.176 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.28e-01 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.gate_up_proj_weight_quantizers.177 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.14e-01 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.gate_up_proj_weight_quantizers.178 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.gate_up_proj_weight_quantizers.179 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.gate_up_proj_weight_quantizers.180 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.14e-01 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.gate_up_proj_weight_quantizers.181 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.gate_up_proj_weight_quantizers.182 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.19e-01 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.gate_up_proj_weight_quantizers.183 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.50e-01 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.gate_up_proj_weight_quantizers.184 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.82e-01 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.gate_up_proj_weight_quantizers.185 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.gate_up_proj_weight_quantizers.186 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.gate_up_proj_weight_quantizers.187 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.gate_up_proj_weight_quantizers.188 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.gate_up_proj_weight_quantizers.189 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.gate_up_proj_weight_quantizers.190 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.gate_up_proj_weight_quantizers.191 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.37e-01 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.gate_up_proj_weight_quantizers.192 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.gate_up_proj_weight_quantizers.193 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.gate_up_proj_weight_quantizers.194 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.gate_up_proj_weight_quantizers.195 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.gate_up_proj_weight_quantizers.196 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.gate_up_proj_weight_quantizers.197 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.gate_up_proj_weight_quantizers.198 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.gate_up_proj_weight_quantizers.199 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.01e-01 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.gate_up_proj_weight_quantizers.200 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.gate_up_proj_weight_quantizers.201 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.gate_up_proj_weight_quantizers.202 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.gate_up_proj_weight_quantizers.203 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.gate_up_proj_weight_quantizers.204 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.gate_up_proj_weight_quantizers.205 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.gate_up_proj_weight_quantizers.206 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.27e-01 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.gate_up_proj_weight_quantizers.207 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.gate_up_proj_weight_quantizers.208 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.gate_up_proj_weight_quantizers.209 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.gate_up_proj_weight_quantizers.210 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.gate_up_proj_weight_quantizers.211 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.gate_up_proj_weight_quantizers.212 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.gate_up_proj_weight_quantizers.213 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.19e-01 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.gate_up_proj_weight_quantizers.214 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.15e-01 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.gate_up_proj_weight_quantizers.215 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.64e-01 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.gate_up_proj_weight_quantizers.216 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.gate_up_proj_weight_quantizers.217 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.gate_up_proj_weight_quantizers.218 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.gate_up_proj_weight_quantizers.219 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.42e-01 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.gate_up_proj_weight_quantizers.220 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.gate_up_proj_weight_quantizers.221 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.gate_up_proj_weight_quantizers.222 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.gate_up_proj_weight_quantizers.223 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.gate_up_proj_weight_quantizers.224 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.gate_up_proj_weight_quantizers.225 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.gate_up_proj_weight_quantizers.226 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.02e-01 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.gate_up_proj_weight_quantizers.227 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.gate_up_proj_weight_quantizers.228 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.gate_up_proj_weight_quantizers.229 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.29e-01 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.gate_up_proj_weight_quantizers.230 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.gate_up_proj_weight_quantizers.231 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.92e-01 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.gate_up_proj_weight_quantizers.232 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.gate_up_proj_weight_quantizers.233 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.gate_up_proj_weight_quantizers.234 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.gate_up_proj_weight_quantizers.235 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.17e-01 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.gate_up_proj_weight_quantizers.236 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.gate_up_proj_weight_quantizers.237 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.gate_up_proj_weight_quantizers.238 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.gate_up_proj_weight_quantizers.239 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.gate_up_proj_weight_quantizers.240 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.gate_up_proj_weight_quantizers.241 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.gate_up_proj_weight_quantizers.242 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.gate_up_proj_weight_quantizers.243 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.gate_up_proj_weight_quantizers.244 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.17e-01 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.gate_up_proj_weight_quantizers.245 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.gate_up_proj_weight_quantizers.246 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.gate_up_proj_weight_quantizers.247 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.16e-01 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.gate_up_proj_weight_quantizers.248 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.gate_up_proj_weight_quantizers.249 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.18e-01 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.gate_up_proj_weight_quantizers.250 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.gate_up_proj_weight_quantizers.251 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.gate_up_proj_weight_quantizers.252 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.gate_up_proj_weight_quantizers.253 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.14e-01 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.gate_up_proj_weight_quantizers.254 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.gate_up_proj_weight_quantizers.255 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.down_proj_input_quantizer TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=5.42e+01 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.down_proj_weight_quantizers.0 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.down_proj_weight_quantizers.1 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.down_proj_weight_quantizers.2 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.down_proj_weight_quantizers.3 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.down_proj_weight_quantizers.4 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.down_proj_weight_quantizers.5 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.down_proj_weight_quantizers.6 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.down_proj_weight_quantizers.7 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.down_proj_weight_quantizers.8 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.down_proj_weight_quantizers.9 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.down_proj_weight_quantizers.10 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.down_proj_weight_quantizers.11 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.down_proj_weight_quantizers.12 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.down_proj_weight_quantizers.13 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.down_proj_weight_quantizers.14 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.down_proj_weight_quantizers.15 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.down_proj_weight_quantizers.16 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.down_proj_weight_quantizers.17 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.down_proj_weight_quantizers.18 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.down_proj_weight_quantizers.19 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.down_proj_weight_quantizers.20 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.down_proj_weight_quantizers.21 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.down_proj_weight_quantizers.22 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.down_proj_weight_quantizers.23 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.down_proj_weight_quantizers.24 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.down_proj_weight_quantizers.25 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.down_proj_weight_quantizers.26 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.down_proj_weight_quantizers.27 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.down_proj_weight_quantizers.28 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.down_proj_weight_quantizers.29 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.down_proj_weight_quantizers.30 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.down_proj_weight_quantizers.31 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.down_proj_weight_quantizers.32 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.down_proj_weight_quantizers.33 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.down_proj_weight_quantizers.34 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.down_proj_weight_quantizers.35 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.down_proj_weight_quantizers.36 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.51e-01 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.down_proj_weight_quantizers.37 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.down_proj_weight_quantizers.38 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.down_proj_weight_quantizers.39 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.75e-01 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.down_proj_weight_quantizers.40 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.down_proj_weight_quantizers.41 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.down_proj_weight_quantizers.42 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.73e-01 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.down_proj_weight_quantizers.43 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.48e-01 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.down_proj_weight_quantizers.44 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.down_proj_weight_quantizers.45 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.down_proj_weight_quantizers.46 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.down_proj_weight_quantizers.47 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.down_proj_weight_quantizers.48 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.down_proj_weight_quantizers.49 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.down_proj_weight_quantizers.50 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.19e-01 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.down_proj_weight_quantizers.51 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.down_proj_weight_quantizers.52 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.down_proj_weight_quantizers.53 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.down_proj_weight_quantizers.54 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.down_proj_weight_quantizers.55 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.down_proj_weight_quantizers.56 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.down_proj_weight_quantizers.57 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.65e-01 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.down_proj_weight_quantizers.58 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.down_proj_weight_quantizers.59 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.down_proj_weight_quantizers.60 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.down_proj_weight_quantizers.61 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.down_proj_weight_quantizers.62 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.down_proj_weight_quantizers.63 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.down_proj_weight_quantizers.64 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.down_proj_weight_quantizers.65 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.down_proj_weight_quantizers.66 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.down_proj_weight_quantizers.67 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.down_proj_weight_quantizers.68 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.down_proj_weight_quantizers.69 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.50e-01 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.down_proj_weight_quantizers.70 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.down_proj_weight_quantizers.71 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.down_proj_weight_quantizers.72 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.down_proj_weight_quantizers.73 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.down_proj_weight_quantizers.74 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.down_proj_weight_quantizers.75 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.down_proj_weight_quantizers.76 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.down_proj_weight_quantizers.77 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.down_proj_weight_quantizers.78 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.58e-01 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.down_proj_weight_quantizers.79 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.down_proj_weight_quantizers.80 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.59e-02 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.down_proj_weight_quantizers.81 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.40e-02 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.down_proj_weight_quantizers.82 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.down_proj_weight_quantizers.83 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.down_proj_weight_quantizers.84 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.down_proj_weight_quantizers.85 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.25e-01 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.down_proj_weight_quantizers.86 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.down_proj_weight_quantizers.87 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.down_proj_weight_quantizers.88 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.down_proj_weight_quantizers.89 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.46e-01 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.down_proj_weight_quantizers.90 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.down_proj_weight_quantizers.91 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.down_proj_weight_quantizers.92 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.down_proj_weight_quantizers.93 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.down_proj_weight_quantizers.94 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.down_proj_weight_quantizers.95 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.down_proj_weight_quantizers.96 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.down_proj_weight_quantizers.97 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.down_proj_weight_quantizers.98 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.down_proj_weight_quantizers.99 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.down_proj_weight_quantizers.100 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.down_proj_weight_quantizers.101 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.down_proj_weight_quantizers.102 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.down_proj_weight_quantizers.103 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.25e-01 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.down_proj_weight_quantizers.104 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.down_proj_weight_quantizers.105 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.down_proj_weight_quantizers.106 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.down_proj_weight_quantizers.107 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.down_proj_weight_quantizers.108 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.down_proj_weight_quantizers.109 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.down_proj_weight_quantizers.110 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.down_proj_weight_quantizers.111 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.down_proj_weight_quantizers.112 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.23e-01 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.down_proj_weight_quantizers.113 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.down_proj_weight_quantizers.114 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.down_proj_weight_quantizers.115 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.down_proj_weight_quantizers.116 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.59e-02 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.down_proj_weight_quantizers.117 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.down_proj_weight_quantizers.118 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.64e-02 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.down_proj_weight_quantizers.119 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.37e-01 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.down_proj_weight_quantizers.120 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.down_proj_weight_quantizers.121 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.down_proj_weight_quantizers.122 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.54e-02 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.down_proj_weight_quantizers.123 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.down_proj_weight_quantizers.124 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.down_proj_weight_quantizers.125 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.down_proj_weight_quantizers.126 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.down_proj_weight_quantizers.127 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.59e-02 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.down_proj_weight_quantizers.128 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.down_proj_weight_quantizers.129 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.down_proj_weight_quantizers.130 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.down_proj_weight_quantizers.131 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.down_proj_weight_quantizers.132 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.63e-01 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.down_proj_weight_quantizers.133 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.down_proj_weight_quantizers.134 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.down_proj_weight_quantizers.135 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.down_proj_weight_quantizers.136 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.down_proj_weight_quantizers.137 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.down_proj_weight_quantizers.138 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.down_proj_weight_quantizers.139 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.down_proj_weight_quantizers.140 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.down_proj_weight_quantizers.141 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.down_proj_weight_quantizers.142 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.64e-02 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.down_proj_weight_quantizers.143 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.down_proj_weight_quantizers.144 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.down_proj_weight_quantizers.145 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.down_proj_weight_quantizers.146 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.down_proj_weight_quantizers.147 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.down_proj_weight_quantizers.148 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.down_proj_weight_quantizers.149 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.down_proj_weight_quantizers.150 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.56e-01 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.down_proj_weight_quantizers.151 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.64e-01 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.down_proj_weight_quantizers.152 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.down_proj_weight_quantizers.153 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.down_proj_weight_quantizers.154 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.down_proj_weight_quantizers.155 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.down_proj_weight_quantizers.156 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.down_proj_weight_quantizers.157 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.down_proj_weight_quantizers.158 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.down_proj_weight_quantizers.159 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.down_proj_weight_quantizers.160 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.down_proj_weight_quantizers.161 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.down_proj_weight_quantizers.162 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.down_proj_weight_quantizers.163 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.down_proj_weight_quantizers.164 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.down_proj_weight_quantizers.165 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.down_proj_weight_quantizers.166 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.down_proj_weight_quantizers.167 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.down_proj_weight_quantizers.168 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.59e-02 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.down_proj_weight_quantizers.169 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.61e-01 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.down_proj_weight_quantizers.170 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.13e-01 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.down_proj_weight_quantizers.171 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.down_proj_weight_quantizers.172 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.down_proj_weight_quantizers.173 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.03e-01 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.down_proj_weight_quantizers.174 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.down_proj_weight_quantizers.175 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.50e-02 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.down_proj_weight_quantizers.176 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.down_proj_weight_quantizers.177 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.down_proj_weight_quantizers.178 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.down_proj_weight_quantizers.179 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.down_proj_weight_quantizers.180 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.down_proj_weight_quantizers.181 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.down_proj_weight_quantizers.182 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.down_proj_weight_quantizers.183 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.down_proj_weight_quantizers.184 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.down_proj_weight_quantizers.185 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.down_proj_weight_quantizers.186 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.down_proj_weight_quantizers.187 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.down_proj_weight_quantizers.188 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.down_proj_weight_quantizers.189 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.down_proj_weight_quantizers.190 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.down_proj_weight_quantizers.191 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.21e-01 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.down_proj_weight_quantizers.192 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.down_proj_weight_quantizers.193 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.down_proj_weight_quantizers.194 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.39e-01 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.down_proj_weight_quantizers.195 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.down_proj_weight_quantizers.196 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.down_proj_weight_quantizers.197 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.down_proj_weight_quantizers.198 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.down_proj_weight_quantizers.199 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.32e-01 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.down_proj_weight_quantizers.200 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.down_proj_weight_quantizers.201 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.down_proj_weight_quantizers.202 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.down_proj_weight_quantizers.203 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.down_proj_weight_quantizers.204 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.down_proj_weight_quantizers.205 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.14e-01 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.down_proj_weight_quantizers.206 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.down_proj_weight_quantizers.207 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.down_proj_weight_quantizers.208 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.down_proj_weight_quantizers.209 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.down_proj_weight_quantizers.210 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.down_proj_weight_quantizers.211 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.25e-02 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.down_proj_weight_quantizers.212 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.down_proj_weight_quantizers.213 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.down_proj_weight_quantizers.214 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.down_proj_weight_quantizers.215 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.71e-01 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.down_proj_weight_quantizers.216 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.64e-02 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.down_proj_weight_quantizers.217 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.down_proj_weight_quantizers.218 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.32e-01 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.down_proj_weight_quantizers.219 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.45e-02 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.down_proj_weight_quantizers.220 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.down_proj_weight_quantizers.221 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.down_proj_weight_quantizers.222 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.down_proj_weight_quantizers.223 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.29e-01 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.down_proj_weight_quantizers.224 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.down_proj_weight_quantizers.225 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.down_proj_weight_quantizers.226 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.down_proj_weight_quantizers.227 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.87e-01 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.down_proj_weight_quantizers.228 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.down_proj_weight_quantizers.229 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.down_proj_weight_quantizers.230 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.down_proj_weight_quantizers.231 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.46e-01 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.down_proj_weight_quantizers.232 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.down_proj_weight_quantizers.233 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.down_proj_weight_quantizers.234 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.down_proj_weight_quantizers.235 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.down_proj_weight_quantizers.236 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.52e-01 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.down_proj_weight_quantizers.237 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.down_proj_weight_quantizers.238 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.down_proj_weight_quantizers.239 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.down_proj_weight_quantizers.240 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.down_proj_weight_quantizers.241 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.down_proj_weight_quantizers.242 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.20e-02 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.down_proj_weight_quantizers.243 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.down_proj_weight_quantizers.244 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.15e-01 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.down_proj_weight_quantizers.245 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.down_proj_weight_quantizers.246 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.down_proj_weight_quantizers.247 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.down_proj_weight_quantizers.248 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.down_proj_weight_quantizers.249 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.76e-01 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.down_proj_weight_quantizers.250 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.down_proj_weight_quantizers.251 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.45e-02 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.down_proj_weight_quantizers.252 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.down_proj_weight_quantizers.253 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.down_proj_weight_quantizers.254 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.down_proj_weight_quantizers.255 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.77e-01 calibrator=MaxCalibrator quant) +model.layers.59.mlp.shared_experts.gate_proj.input_quantizer TensorQuantizer(disabled) +model.layers.59.mlp.shared_experts.gate_proj.output_quantizer TensorQuantizer(disabled) +model.layers.59.mlp.shared_experts.gate_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.59.mlp.shared_experts.up_proj.input_quantizer TensorQuantizer(disabled) +model.layers.59.mlp.shared_experts.up_proj.output_quantizer TensorQuantizer(disabled) +model.layers.59.mlp.shared_experts.up_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.59.mlp.shared_experts.down_proj.input_quantizer TensorQuantizer(disabled) +model.layers.59.mlp.shared_experts.down_proj.output_quantizer TensorQuantizer(disabled) +model.layers.59.mlp.shared_experts.down_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.60.self_attn.q_a_proj.input_quantizer TensorQuantizer(disabled) +model.layers.60.self_attn.q_a_proj.output_quantizer TensorQuantizer(disabled) +model.layers.60.self_attn.q_a_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.60.self_attn.q_b_proj.input_quantizer TensorQuantizer(disabled) +model.layers.60.self_attn.q_b_proj.output_quantizer TensorQuantizer(disabled) +model.layers.60.self_attn.q_b_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.60.self_attn.kv_a_proj_with_mqa.input_quantizer TensorQuantizer(disabled) +model.layers.60.self_attn.kv_a_proj_with_mqa.output_quantizer TensorQuantizer(disabled) +model.layers.60.self_attn.kv_a_proj_with_mqa.weight_quantizer TensorQuantizer(disabled) +model.layers.60.self_attn.kv_b_proj.input_quantizer TensorQuantizer(disabled) +model.layers.60.self_attn.kv_b_proj.output_quantizer TensorQuantizer(disabled) +model.layers.60.self_attn.kv_b_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.60.self_attn.o_proj.input_quantizer TensorQuantizer(disabled) +model.layers.60.self_attn.o_proj.output_quantizer TensorQuantizer(disabled) +model.layers.60.self_attn.o_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.60.self_attn.q_bmm_quantizer TensorQuantizer(disabled) +model.layers.60.self_attn.k_bmm_quantizer TensorQuantizer((4, 3) bit fake per-tensor amax=4.48e+02(const) calibrator=MaxCalibrator quant) +model.layers.60.self_attn.v_bmm_quantizer TensorQuantizer((4, 3) bit fake per-tensor amax=4.48e+02(const) calibrator=MaxCalibrator quant) +model.layers.60.self_attn.p_bmm_quantizer TensorQuantizer(disabled) +model.layers.60.mlp.experts.gate_up_proj_input_quantizer TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.81e+00 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.gate_up_proj_weight_quantizers.0 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.42e-01 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.gate_up_proj_weight_quantizers.1 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.17e-01 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.gate_up_proj_weight_quantizers.2 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.gate_up_proj_weight_quantizers.3 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.gate_up_proj_weight_quantizers.4 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.gate_up_proj_weight_quantizers.5 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.14e-01 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.gate_up_proj_weight_quantizers.6 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.gate_up_proj_weight_quantizers.7 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.gate_up_proj_weight_quantizers.8 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.gate_up_proj_weight_quantizers.9 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.gate_up_proj_weight_quantizers.10 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.gate_up_proj_weight_quantizers.11 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.gate_up_proj_weight_quantizers.12 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.23e-01 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.gate_up_proj_weight_quantizers.13 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.gate_up_proj_weight_quantizers.14 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.gate_up_proj_weight_quantizers.15 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.gate_up_proj_weight_quantizers.16 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.gate_up_proj_weight_quantizers.17 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.gate_up_proj_weight_quantizers.18 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.gate_up_proj_weight_quantizers.19 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.53e-01 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.gate_up_proj_weight_quantizers.20 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.gate_up_proj_weight_quantizers.21 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.gate_up_proj_weight_quantizers.22 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.14e-01 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.gate_up_proj_weight_quantizers.23 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.gate_up_proj_weight_quantizers.24 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.gate_up_proj_weight_quantizers.25 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.60e-01 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.gate_up_proj_weight_quantizers.26 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.gate_up_proj_weight_quantizers.27 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.gate_up_proj_weight_quantizers.28 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.gate_up_proj_weight_quantizers.29 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.81e-01 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.gate_up_proj_weight_quantizers.30 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.gate_up_proj_weight_quantizers.31 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.gate_up_proj_weight_quantizers.32 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.gate_up_proj_weight_quantizers.33 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.gate_up_proj_weight_quantizers.34 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.gate_up_proj_weight_quantizers.35 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.40e-01 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.gate_up_proj_weight_quantizers.36 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.gate_up_proj_weight_quantizers.37 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.gate_up_proj_weight_quantizers.38 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.76e-01 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.gate_up_proj_weight_quantizers.39 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.17e-01 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.gate_up_proj_weight_quantizers.40 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.gate_up_proj_weight_quantizers.41 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.61e-01 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.gate_up_proj_weight_quantizers.42 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.gate_up_proj_weight_quantizers.43 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.gate_up_proj_weight_quantizers.44 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.gate_up_proj_weight_quantizers.45 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.85e-01 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.gate_up_proj_weight_quantizers.46 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.gate_up_proj_weight_quantizers.47 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.gate_up_proj_weight_quantizers.48 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.gate_up_proj_weight_quantizers.49 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.76e-01 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.gate_up_proj_weight_quantizers.50 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.gate_up_proj_weight_quantizers.51 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.gate_up_proj_weight_quantizers.52 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.gate_up_proj_weight_quantizers.53 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.gate_up_proj_weight_quantizers.54 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.28e-01 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.gate_up_proj_weight_quantizers.55 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.24e-01 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.gate_up_proj_weight_quantizers.56 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.gate_up_proj_weight_quantizers.57 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.gate_up_proj_weight_quantizers.58 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.gate_up_proj_weight_quantizers.59 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.gate_up_proj_weight_quantizers.60 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.gate_up_proj_weight_quantizers.61 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.gate_up_proj_weight_quantizers.62 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.gate_up_proj_weight_quantizers.63 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.gate_up_proj_weight_quantizers.64 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.gate_up_proj_weight_quantizers.65 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.gate_up_proj_weight_quantizers.66 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.gate_up_proj_weight_quantizers.67 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.gate_up_proj_weight_quantizers.68 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.gate_up_proj_weight_quantizers.69 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.gate_up_proj_weight_quantizers.70 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.25e-01 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.gate_up_proj_weight_quantizers.71 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.gate_up_proj_weight_quantizers.72 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.gate_up_proj_weight_quantizers.73 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.gate_up_proj_weight_quantizers.74 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.42e-01 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.gate_up_proj_weight_quantizers.75 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.gate_up_proj_weight_quantizers.76 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.gate_up_proj_weight_quantizers.77 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.gate_up_proj_weight_quantizers.78 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.50e-01 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.gate_up_proj_weight_quantizers.79 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.44e-01 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.gate_up_proj_weight_quantizers.80 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.gate_up_proj_weight_quantizers.81 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.gate_up_proj_weight_quantizers.82 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.31e-01 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.gate_up_proj_weight_quantizers.83 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.gate_up_proj_weight_quantizers.84 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.gate_up_proj_weight_quantizers.85 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.74e-01 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.gate_up_proj_weight_quantizers.86 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.gate_up_proj_weight_quantizers.87 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.gate_up_proj_weight_quantizers.88 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.gate_up_proj_weight_quantizers.89 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.gate_up_proj_weight_quantizers.90 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.gate_up_proj_weight_quantizers.91 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.gate_up_proj_weight_quantizers.92 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.66e-01 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.gate_up_proj_weight_quantizers.93 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.gate_up_proj_weight_quantizers.94 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.gate_up_proj_weight_quantizers.95 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.13e-01 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.gate_up_proj_weight_quantizers.96 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.gate_up_proj_weight_quantizers.97 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.gate_up_proj_weight_quantizers.98 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.gate_up_proj_weight_quantizers.99 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.gate_up_proj_weight_quantizers.100 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.13e-01 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.gate_up_proj_weight_quantizers.101 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.gate_up_proj_weight_quantizers.102 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.gate_up_proj_weight_quantizers.103 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.gate_up_proj_weight_quantizers.104 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.gate_up_proj_weight_quantizers.105 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.17e-01 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.gate_up_proj_weight_quantizers.106 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.gate_up_proj_weight_quantizers.107 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.gate_up_proj_weight_quantizers.108 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.gate_up_proj_weight_quantizers.109 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.48e-01 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.gate_up_proj_weight_quantizers.110 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.19e-01 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.gate_up_proj_weight_quantizers.111 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.gate_up_proj_weight_quantizers.112 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.23e-01 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.gate_up_proj_weight_quantizers.113 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.gate_up_proj_weight_quantizers.114 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.gate_up_proj_weight_quantizers.115 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.gate_up_proj_weight_quantizers.116 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.gate_up_proj_weight_quantizers.117 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.gate_up_proj_weight_quantizers.118 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.19e-01 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.gate_up_proj_weight_quantizers.119 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.gate_up_proj_weight_quantizers.120 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.gate_up_proj_weight_quantizers.121 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.gate_up_proj_weight_quantizers.122 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.gate_up_proj_weight_quantizers.123 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.gate_up_proj_weight_quantizers.124 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.17e-01 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.gate_up_proj_weight_quantizers.125 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.28e-01 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.gate_up_proj_weight_quantizers.126 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.gate_up_proj_weight_quantizers.127 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.gate_up_proj_weight_quantizers.128 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.gate_up_proj_weight_quantizers.129 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.20e-01 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.gate_up_proj_weight_quantizers.130 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.gate_up_proj_weight_quantizers.131 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.gate_up_proj_weight_quantizers.132 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.gate_up_proj_weight_quantizers.133 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.gate_up_proj_weight_quantizers.134 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.gate_up_proj_weight_quantizers.135 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.gate_up_proj_weight_quantizers.136 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.gate_up_proj_weight_quantizers.137 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.gate_up_proj_weight_quantizers.138 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.gate_up_proj_weight_quantizers.139 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.18e-01 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.gate_up_proj_weight_quantizers.140 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.gate_up_proj_weight_quantizers.141 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.gate_up_proj_weight_quantizers.142 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.gate_up_proj_weight_quantizers.143 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.gate_up_proj_weight_quantizers.144 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.gate_up_proj_weight_quantizers.145 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.gate_up_proj_weight_quantizers.146 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.53e-01 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.gate_up_proj_weight_quantizers.147 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.gate_up_proj_weight_quantizers.148 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.40e-01 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.gate_up_proj_weight_quantizers.149 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.17e-01 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.gate_up_proj_weight_quantizers.150 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.gate_up_proj_weight_quantizers.151 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.gate_up_proj_weight_quantizers.152 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.18e-01 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.gate_up_proj_weight_quantizers.153 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.31e-01 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.gate_up_proj_weight_quantizers.154 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.gate_up_proj_weight_quantizers.155 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.16e-01 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.gate_up_proj_weight_quantizers.156 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.gate_up_proj_weight_quantizers.157 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.gate_up_proj_weight_quantizers.158 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.gate_up_proj_weight_quantizers.159 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.gate_up_proj_weight_quantizers.160 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.gate_up_proj_weight_quantizers.161 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.gate_up_proj_weight_quantizers.162 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.gate_up_proj_weight_quantizers.163 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.gate_up_proj_weight_quantizers.164 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.gate_up_proj_weight_quantizers.165 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.gate_up_proj_weight_quantizers.166 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.gate_up_proj_weight_quantizers.167 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.gate_up_proj_weight_quantizers.168 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.gate_up_proj_weight_quantizers.169 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.gate_up_proj_weight_quantizers.170 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.gate_up_proj_weight_quantizers.171 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.70e-01 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.gate_up_proj_weight_quantizers.172 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.gate_up_proj_weight_quantizers.173 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.46e-01 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.gate_up_proj_weight_quantizers.174 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.gate_up_proj_weight_quantizers.175 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.28e-01 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.gate_up_proj_weight_quantizers.176 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.gate_up_proj_weight_quantizers.177 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.13e-01 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.gate_up_proj_weight_quantizers.178 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.gate_up_proj_weight_quantizers.179 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.gate_up_proj_weight_quantizers.180 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.gate_up_proj_weight_quantizers.181 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.gate_up_proj_weight_quantizers.182 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.gate_up_proj_weight_quantizers.183 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.gate_up_proj_weight_quantizers.184 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.19e-01 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.gate_up_proj_weight_quantizers.185 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.gate_up_proj_weight_quantizers.186 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.gate_up_proj_weight_quantizers.187 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.gate_up_proj_weight_quantizers.188 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.29e-01 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.gate_up_proj_weight_quantizers.189 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.gate_up_proj_weight_quantizers.190 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.gate_up_proj_weight_quantizers.191 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.gate_up_proj_weight_quantizers.192 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.56e-01 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.gate_up_proj_weight_quantizers.193 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.gate_up_proj_weight_quantizers.194 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.gate_up_proj_weight_quantizers.195 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.gate_up_proj_weight_quantizers.196 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.gate_up_proj_weight_quantizers.197 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.15e-01 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.gate_up_proj_weight_quantizers.198 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.gate_up_proj_weight_quantizers.199 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.gate_up_proj_weight_quantizers.200 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.62e-01 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.gate_up_proj_weight_quantizers.201 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.85e-01 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.gate_up_proj_weight_quantizers.202 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.gate_up_proj_weight_quantizers.203 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.gate_up_proj_weight_quantizers.204 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.75e-01 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.gate_up_proj_weight_quantizers.205 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.gate_up_proj_weight_quantizers.206 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.gate_up_proj_weight_quantizers.207 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.gate_up_proj_weight_quantizers.208 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.gate_up_proj_weight_quantizers.209 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.16e-01 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.gate_up_proj_weight_quantizers.210 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.19e-01 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.gate_up_proj_weight_quantizers.211 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.gate_up_proj_weight_quantizers.212 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.gate_up_proj_weight_quantizers.213 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.gate_up_proj_weight_quantizers.214 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.gate_up_proj_weight_quantizers.215 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.gate_up_proj_weight_quantizers.216 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.gate_up_proj_weight_quantizers.217 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.gate_up_proj_weight_quantizers.218 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.gate_up_proj_weight_quantizers.219 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.gate_up_proj_weight_quantizers.220 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.gate_up_proj_weight_quantizers.221 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.gate_up_proj_weight_quantizers.222 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.gate_up_proj_weight_quantizers.223 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.gate_up_proj_weight_quantizers.224 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.gate_up_proj_weight_quantizers.225 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.gate_up_proj_weight_quantizers.226 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.gate_up_proj_weight_quantizers.227 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.gate_up_proj_weight_quantizers.228 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.gate_up_proj_weight_quantizers.229 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.17e-01 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.gate_up_proj_weight_quantizers.230 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.gate_up_proj_weight_quantizers.231 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.gate_up_proj_weight_quantizers.232 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.gate_up_proj_weight_quantizers.233 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.68e-01 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.gate_up_proj_weight_quantizers.234 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.gate_up_proj_weight_quantizers.235 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.gate_up_proj_weight_quantizers.236 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.gate_up_proj_weight_quantizers.237 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.gate_up_proj_weight_quantizers.238 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.gate_up_proj_weight_quantizers.239 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.gate_up_proj_weight_quantizers.240 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.gate_up_proj_weight_quantizers.241 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.gate_up_proj_weight_quantizers.242 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.gate_up_proj_weight_quantizers.243 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.gate_up_proj_weight_quantizers.244 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.gate_up_proj_weight_quantizers.245 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.gate_up_proj_weight_quantizers.246 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.16e-01 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.gate_up_proj_weight_quantizers.247 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.gate_up_proj_weight_quantizers.248 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.gate_up_proj_weight_quantizers.249 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.gate_up_proj_weight_quantizers.250 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.gate_up_proj_weight_quantizers.251 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.48e-01 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.gate_up_proj_weight_quantizers.252 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.gate_up_proj_weight_quantizers.253 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.gate_up_proj_weight_quantizers.254 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.gate_up_proj_weight_quantizers.255 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.29e-01 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.down_proj_input_quantizer TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.08e+01 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.down_proj_weight_quantizers.0 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.62e-01 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.down_proj_weight_quantizers.1 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.down_proj_weight_quantizers.2 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.down_proj_weight_quantizers.3 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.down_proj_weight_quantizers.4 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.down_proj_weight_quantizers.5 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.down_proj_weight_quantizers.6 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.down_proj_weight_quantizers.7 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.down_proj_weight_quantizers.8 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.down_proj_weight_quantizers.9 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.33e-01 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.down_proj_weight_quantizers.10 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.down_proj_weight_quantizers.11 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.down_proj_weight_quantizers.12 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.down_proj_weight_quantizers.13 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.down_proj_weight_quantizers.14 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.down_proj_weight_quantizers.15 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.down_proj_weight_quantizers.16 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.down_proj_weight_quantizers.17 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.down_proj_weight_quantizers.18 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.59e-01 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.down_proj_weight_quantizers.19 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.down_proj_weight_quantizers.20 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.down_proj_weight_quantizers.21 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.down_proj_weight_quantizers.22 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.down_proj_weight_quantizers.23 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.down_proj_weight_quantizers.24 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.down_proj_weight_quantizers.25 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.50e-02 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.down_proj_weight_quantizers.26 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.down_proj_weight_quantizers.27 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.down_proj_weight_quantizers.28 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.67e-01 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.down_proj_weight_quantizers.29 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.down_proj_weight_quantizers.30 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.down_proj_weight_quantizers.31 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.down_proj_weight_quantizers.32 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.down_proj_weight_quantizers.33 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.down_proj_weight_quantizers.34 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.64e-02 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.down_proj_weight_quantizers.35 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.91e-01 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.down_proj_weight_quantizers.36 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.47e-01 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.down_proj_weight_quantizers.37 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.19e-01 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.down_proj_weight_quantizers.38 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.down_proj_weight_quantizers.39 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.down_proj_weight_quantizers.40 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.down_proj_weight_quantizers.41 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.50e-02 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.down_proj_weight_quantizers.42 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.down_proj_weight_quantizers.43 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.down_proj_weight_quantizers.44 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.down_proj_weight_quantizers.45 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.00e-01 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.down_proj_weight_quantizers.46 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.down_proj_weight_quantizers.47 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.down_proj_weight_quantizers.48 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.down_proj_weight_quantizers.49 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.down_proj_weight_quantizers.50 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.down_proj_weight_quantizers.51 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.down_proj_weight_quantizers.52 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.down_proj_weight_quantizers.53 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.47e-01 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.down_proj_weight_quantizers.54 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.25e-01 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.down_proj_weight_quantizers.55 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.down_proj_weight_quantizers.56 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.down_proj_weight_quantizers.57 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.down_proj_weight_quantizers.58 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.down_proj_weight_quantizers.59 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.down_proj_weight_quantizers.60 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.down_proj_weight_quantizers.61 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.down_proj_weight_quantizers.62 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.down_proj_weight_quantizers.63 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.49e-01 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.down_proj_weight_quantizers.64 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.down_proj_weight_quantizers.65 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.down_proj_weight_quantizers.66 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.down_proj_weight_quantizers.67 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.59e-02 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.down_proj_weight_quantizers.68 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.down_proj_weight_quantizers.69 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.down_proj_weight_quantizers.70 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.down_proj_weight_quantizers.71 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.down_proj_weight_quantizers.72 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.down_proj_weight_quantizers.73 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.down_proj_weight_quantizers.74 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.down_proj_weight_quantizers.75 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.down_proj_weight_quantizers.76 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.down_proj_weight_quantizers.77 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.down_proj_weight_quantizers.78 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.down_proj_weight_quantizers.79 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.down_proj_weight_quantizers.80 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.down_proj_weight_quantizers.81 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.down_proj_weight_quantizers.82 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.down_proj_weight_quantizers.83 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.down_proj_weight_quantizers.84 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.down_proj_weight_quantizers.85 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.down_proj_weight_quantizers.86 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.down_proj_weight_quantizers.87 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.54e-02 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.down_proj_weight_quantizers.88 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.down_proj_weight_quantizers.89 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.down_proj_weight_quantizers.90 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.down_proj_weight_quantizers.91 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.down_proj_weight_quantizers.92 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.82e-01 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.down_proj_weight_quantizers.93 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.down_proj_weight_quantizers.94 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.down_proj_weight_quantizers.95 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.down_proj_weight_quantizers.96 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.15e-01 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.down_proj_weight_quantizers.97 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.45e-02 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.down_proj_weight_quantizers.98 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.down_proj_weight_quantizers.99 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.down_proj_weight_quantizers.100 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.down_proj_weight_quantizers.101 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.down_proj_weight_quantizers.102 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.down_proj_weight_quantizers.103 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.down_proj_weight_quantizers.104 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.down_proj_weight_quantizers.105 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.61e-01 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.down_proj_weight_quantizers.106 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.down_proj_weight_quantizers.107 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.down_proj_weight_quantizers.108 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.down_proj_weight_quantizers.109 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.down_proj_weight_quantizers.110 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.down_proj_weight_quantizers.111 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.down_proj_weight_quantizers.112 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.down_proj_weight_quantizers.113 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.down_proj_weight_quantizers.114 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.down_proj_weight_quantizers.115 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.59e-02 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.down_proj_weight_quantizers.116 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.down_proj_weight_quantizers.117 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.down_proj_weight_quantizers.118 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.96e-01 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.down_proj_weight_quantizers.119 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.down_proj_weight_quantizers.120 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.down_proj_weight_quantizers.121 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.down_proj_weight_quantizers.122 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.64e-02 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.down_proj_weight_quantizers.123 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.down_proj_weight_quantizers.124 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.13e-01 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.down_proj_weight_quantizers.125 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.down_proj_weight_quantizers.126 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.down_proj_weight_quantizers.127 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.15e-01 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.down_proj_weight_quantizers.128 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.down_proj_weight_quantizers.129 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.down_proj_weight_quantizers.130 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.down_proj_weight_quantizers.131 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.down_proj_weight_quantizers.132 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.down_proj_weight_quantizers.133 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.down_proj_weight_quantizers.134 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.down_proj_weight_quantizers.135 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.down_proj_weight_quantizers.136 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.down_proj_weight_quantizers.137 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.down_proj_weight_quantizers.138 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.down_proj_weight_quantizers.139 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.down_proj_weight_quantizers.140 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.down_proj_weight_quantizers.141 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.down_proj_weight_quantizers.142 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.down_proj_weight_quantizers.143 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.down_proj_weight_quantizers.144 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.24e-01 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.down_proj_weight_quantizers.145 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.down_proj_weight_quantizers.146 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.down_proj_weight_quantizers.147 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.down_proj_weight_quantizers.148 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.82e-01 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.down_proj_weight_quantizers.149 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.down_proj_weight_quantizers.150 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.down_proj_weight_quantizers.151 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.down_proj_weight_quantizers.152 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.down_proj_weight_quantizers.153 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.down_proj_weight_quantizers.154 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.down_proj_weight_quantizers.155 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.down_proj_weight_quantizers.156 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.down_proj_weight_quantizers.157 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.down_proj_weight_quantizers.158 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.down_proj_weight_quantizers.159 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.down_proj_weight_quantizers.160 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.down_proj_weight_quantizers.161 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.down_proj_weight_quantizers.162 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.down_proj_weight_quantizers.163 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.down_proj_weight_quantizers.164 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.down_proj_weight_quantizers.165 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.down_proj_weight_quantizers.166 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.down_proj_weight_quantizers.167 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.down_proj_weight_quantizers.168 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.64e-02 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.down_proj_weight_quantizers.169 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.down_proj_weight_quantizers.170 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.down_proj_weight_quantizers.171 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.89e-01 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.down_proj_weight_quantizers.172 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.down_proj_weight_quantizers.173 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.down_proj_weight_quantizers.174 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.down_proj_weight_quantizers.175 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.down_proj_weight_quantizers.176 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.down_proj_weight_quantizers.177 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.down_proj_weight_quantizers.178 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.down_proj_weight_quantizers.179 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.down_proj_weight_quantizers.180 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.down_proj_weight_quantizers.181 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.down_proj_weight_quantizers.182 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.17e-01 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.down_proj_weight_quantizers.183 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.down_proj_weight_quantizers.184 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.down_proj_weight_quantizers.185 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.59e-02 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.down_proj_weight_quantizers.186 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.down_proj_weight_quantizers.187 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.down_proj_weight_quantizers.188 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.down_proj_weight_quantizers.189 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.down_proj_weight_quantizers.190 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.down_proj_weight_quantizers.191 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.down_proj_weight_quantizers.192 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.30e-01 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.down_proj_weight_quantizers.193 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.down_proj_weight_quantizers.194 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.down_proj_weight_quantizers.195 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.54e-02 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.down_proj_weight_quantizers.196 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.down_proj_weight_quantizers.197 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.down_proj_weight_quantizers.198 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.16e-01 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.down_proj_weight_quantizers.199 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.down_proj_weight_quantizers.200 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.down_proj_weight_quantizers.201 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.down_proj_weight_quantizers.202 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.down_proj_weight_quantizers.203 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.down_proj_weight_quantizers.204 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.down_proj_weight_quantizers.205 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.down_proj_weight_quantizers.206 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.down_proj_weight_quantizers.207 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.down_proj_weight_quantizers.208 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.down_proj_weight_quantizers.209 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.83e-01 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.down_proj_weight_quantizers.210 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.down_proj_weight_quantizers.211 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.66e-01 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.down_proj_weight_quantizers.212 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.down_proj_weight_quantizers.213 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.down_proj_weight_quantizers.214 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.35e-01 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.down_proj_weight_quantizers.215 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.down_proj_weight_quantizers.216 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.64e-02 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.down_proj_weight_quantizers.217 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.down_proj_weight_quantizers.218 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.20e-02 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.down_proj_weight_quantizers.219 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.down_proj_weight_quantizers.220 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.down_proj_weight_quantizers.221 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.down_proj_weight_quantizers.222 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.down_proj_weight_quantizers.223 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.down_proj_weight_quantizers.224 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.down_proj_weight_quantizers.225 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.down_proj_weight_quantizers.226 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.down_proj_weight_quantizers.227 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.down_proj_weight_quantizers.228 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.down_proj_weight_quantizers.229 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.down_proj_weight_quantizers.230 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.down_proj_weight_quantizers.231 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.down_proj_weight_quantizers.232 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.down_proj_weight_quantizers.233 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.down_proj_weight_quantizers.234 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.down_proj_weight_quantizers.235 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.down_proj_weight_quantizers.236 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.43e-01 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.down_proj_weight_quantizers.237 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.down_proj_weight_quantizers.238 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.down_proj_weight_quantizers.239 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.down_proj_weight_quantizers.240 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.down_proj_weight_quantizers.241 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.down_proj_weight_quantizers.242 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.down_proj_weight_quantizers.243 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.down_proj_weight_quantizers.244 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.down_proj_weight_quantizers.245 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.down_proj_weight_quantizers.246 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.down_proj_weight_quantizers.247 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.down_proj_weight_quantizers.248 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.down_proj_weight_quantizers.249 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.down_proj_weight_quantizers.250 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.down_proj_weight_quantizers.251 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.down_proj_weight_quantizers.252 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.down_proj_weight_quantizers.253 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.down_proj_weight_quantizers.254 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.down_proj_weight_quantizers.255 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.48e-01 calibrator=MaxCalibrator quant) +model.layers.60.mlp.shared_experts.gate_proj.input_quantizer TensorQuantizer(disabled) +model.layers.60.mlp.shared_experts.gate_proj.output_quantizer TensorQuantizer(disabled) +model.layers.60.mlp.shared_experts.gate_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.60.mlp.shared_experts.up_proj.input_quantizer TensorQuantizer(disabled) +model.layers.60.mlp.shared_experts.up_proj.output_quantizer TensorQuantizer(disabled) +model.layers.60.mlp.shared_experts.up_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.60.mlp.shared_experts.down_proj.input_quantizer TensorQuantizer(disabled) +model.layers.60.mlp.shared_experts.down_proj.output_quantizer TensorQuantizer(disabled) +model.layers.60.mlp.shared_experts.down_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.61.self_attn.q_a_proj.input_quantizer TensorQuantizer(disabled) +model.layers.61.self_attn.q_a_proj.output_quantizer TensorQuantizer(disabled) +model.layers.61.self_attn.q_a_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.61.self_attn.q_b_proj.input_quantizer TensorQuantizer(disabled) +model.layers.61.self_attn.q_b_proj.output_quantizer TensorQuantizer(disabled) +model.layers.61.self_attn.q_b_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.61.self_attn.kv_a_proj_with_mqa.input_quantizer TensorQuantizer(disabled) +model.layers.61.self_attn.kv_a_proj_with_mqa.output_quantizer TensorQuantizer(disabled) +model.layers.61.self_attn.kv_a_proj_with_mqa.weight_quantizer TensorQuantizer(disabled) +model.layers.61.self_attn.kv_b_proj.input_quantizer TensorQuantizer(disabled) +model.layers.61.self_attn.kv_b_proj.output_quantizer TensorQuantizer(disabled) +model.layers.61.self_attn.kv_b_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.61.self_attn.o_proj.input_quantizer TensorQuantizer(disabled) +model.layers.61.self_attn.o_proj.output_quantizer TensorQuantizer(disabled) +model.layers.61.self_attn.o_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.61.self_attn.q_bmm_quantizer TensorQuantizer(disabled) +model.layers.61.self_attn.k_bmm_quantizer TensorQuantizer((4, 3) bit fake per-tensor amax=4.48e+02(const) calibrator=MaxCalibrator quant) +model.layers.61.self_attn.v_bmm_quantizer TensorQuantizer((4, 3) bit fake per-tensor amax=4.48e+02(const) calibrator=MaxCalibrator quant) +model.layers.61.self_attn.p_bmm_quantizer TensorQuantizer(disabled) +model.layers.61.mlp.experts.gate_up_proj_input_quantizer TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.89e+00 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.gate_up_proj_weight_quantizers.0 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.gate_up_proj_weight_quantizers.1 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.15e-01 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.gate_up_proj_weight_quantizers.2 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.27e-01 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.gate_up_proj_weight_quantizers.3 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.37e-01 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.gate_up_proj_weight_quantizers.4 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.28e-01 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.gate_up_proj_weight_quantizers.5 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.92e-01 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.gate_up_proj_weight_quantizers.6 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.gate_up_proj_weight_quantizers.7 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.39e-01 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.gate_up_proj_weight_quantizers.8 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.gate_up_proj_weight_quantizers.9 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.gate_up_proj_weight_quantizers.10 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.gate_up_proj_weight_quantizers.11 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.gate_up_proj_weight_quantizers.12 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.gate_up_proj_weight_quantizers.13 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.gate_up_proj_weight_quantizers.14 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.gate_up_proj_weight_quantizers.15 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.gate_up_proj_weight_quantizers.16 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.gate_up_proj_weight_quantizers.17 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.gate_up_proj_weight_quantizers.18 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.gate_up_proj_weight_quantizers.19 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.59e-01 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.gate_up_proj_weight_quantizers.20 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.16e-01 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.gate_up_proj_weight_quantizers.21 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.gate_up_proj_weight_quantizers.22 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.gate_up_proj_weight_quantizers.23 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.25e-01 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.gate_up_proj_weight_quantizers.24 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.gate_up_proj_weight_quantizers.25 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.37e-01 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.gate_up_proj_weight_quantizers.26 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.gate_up_proj_weight_quantizers.27 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.gate_up_proj_weight_quantizers.28 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.gate_up_proj_weight_quantizers.29 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.gate_up_proj_weight_quantizers.30 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.gate_up_proj_weight_quantizers.31 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.gate_up_proj_weight_quantizers.32 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.gate_up_proj_weight_quantizers.33 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.gate_up_proj_weight_quantizers.34 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.22e-01 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.gate_up_proj_weight_quantizers.35 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.gate_up_proj_weight_quantizers.36 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.gate_up_proj_weight_quantizers.37 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.14e-01 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.gate_up_proj_weight_quantizers.38 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.gate_up_proj_weight_quantizers.39 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.15e-01 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.gate_up_proj_weight_quantizers.40 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.gate_up_proj_weight_quantizers.41 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.gate_up_proj_weight_quantizers.42 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.22e-01 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.gate_up_proj_weight_quantizers.43 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.32e-01 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.gate_up_proj_weight_quantizers.44 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.28e-01 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.gate_up_proj_weight_quantizers.45 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.gate_up_proj_weight_quantizers.46 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.gate_up_proj_weight_quantizers.47 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.gate_up_proj_weight_quantizers.48 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.gate_up_proj_weight_quantizers.49 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.gate_up_proj_weight_quantizers.50 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.gate_up_proj_weight_quantizers.51 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.gate_up_proj_weight_quantizers.52 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.gate_up_proj_weight_quantizers.53 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.gate_up_proj_weight_quantizers.54 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.gate_up_proj_weight_quantizers.55 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.gate_up_proj_weight_quantizers.56 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.23e-01 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.gate_up_proj_weight_quantizers.57 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.gate_up_proj_weight_quantizers.58 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.01e-01 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.gate_up_proj_weight_quantizers.59 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.98e-01 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.gate_up_proj_weight_quantizers.60 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.58e-01 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.gate_up_proj_weight_quantizers.61 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.gate_up_proj_weight_quantizers.62 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.gate_up_proj_weight_quantizers.63 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.80e-01 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.gate_up_proj_weight_quantizers.64 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.gate_up_proj_weight_quantizers.65 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.gate_up_proj_weight_quantizers.66 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.gate_up_proj_weight_quantizers.67 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.gate_up_proj_weight_quantizers.68 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.14e-01 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.gate_up_proj_weight_quantizers.69 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.gate_up_proj_weight_quantizers.70 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.gate_up_proj_weight_quantizers.71 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.gate_up_proj_weight_quantizers.72 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.gate_up_proj_weight_quantizers.73 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.13e-01 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.gate_up_proj_weight_quantizers.74 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.gate_up_proj_weight_quantizers.75 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.gate_up_proj_weight_quantizers.76 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.gate_up_proj_weight_quantizers.77 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.gate_up_proj_weight_quantizers.78 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.26e-01 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.gate_up_proj_weight_quantizers.79 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.gate_up_proj_weight_quantizers.80 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.gate_up_proj_weight_quantizers.81 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.gate_up_proj_weight_quantizers.82 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.gate_up_proj_weight_quantizers.83 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.34e-01 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.gate_up_proj_weight_quantizers.84 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.gate_up_proj_weight_quantizers.85 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.16e-01 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.gate_up_proj_weight_quantizers.86 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.38e-01 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.gate_up_proj_weight_quantizers.87 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.gate_up_proj_weight_quantizers.88 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.gate_up_proj_weight_quantizers.89 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.gate_up_proj_weight_quantizers.90 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.gate_up_proj_weight_quantizers.91 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.97e-01 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.gate_up_proj_weight_quantizers.92 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.gate_up_proj_weight_quantizers.93 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.13e-01 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.gate_up_proj_weight_quantizers.94 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.gate_up_proj_weight_quantizers.95 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.gate_up_proj_weight_quantizers.96 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.gate_up_proj_weight_quantizers.97 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.gate_up_proj_weight_quantizers.98 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.gate_up_proj_weight_quantizers.99 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.gate_up_proj_weight_quantizers.100 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.gate_up_proj_weight_quantizers.101 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.54e-01 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.gate_up_proj_weight_quantizers.102 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.gate_up_proj_weight_quantizers.103 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.gate_up_proj_weight_quantizers.104 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.gate_up_proj_weight_quantizers.105 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.13e-01 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.gate_up_proj_weight_quantizers.106 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.gate_up_proj_weight_quantizers.107 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.gate_up_proj_weight_quantizers.108 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.96e-01 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.gate_up_proj_weight_quantizers.109 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.gate_up_proj_weight_quantizers.110 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.gate_up_proj_weight_quantizers.111 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.gate_up_proj_weight_quantizers.112 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.gate_up_proj_weight_quantizers.113 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.gate_up_proj_weight_quantizers.114 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.gate_up_proj_weight_quantizers.115 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.gate_up_proj_weight_quantizers.116 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.gate_up_proj_weight_quantizers.117 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.gate_up_proj_weight_quantizers.118 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.gate_up_proj_weight_quantizers.119 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.35e-01 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.gate_up_proj_weight_quantizers.120 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.35e-01 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.gate_up_proj_weight_quantizers.121 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.gate_up_proj_weight_quantizers.122 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.54e-02 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.gate_up_proj_weight_quantizers.123 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.gate_up_proj_weight_quantizers.124 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.gate_up_proj_weight_quantizers.125 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.gate_up_proj_weight_quantizers.126 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.gate_up_proj_weight_quantizers.127 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.gate_up_proj_weight_quantizers.128 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.25e-01 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.gate_up_proj_weight_quantizers.129 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.85e-01 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.gate_up_proj_weight_quantizers.130 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.gate_up_proj_weight_quantizers.131 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.gate_up_proj_weight_quantizers.132 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.18e-01 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.gate_up_proj_weight_quantizers.133 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.gate_up_proj_weight_quantizers.134 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.gate_up_proj_weight_quantizers.135 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.gate_up_proj_weight_quantizers.136 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.gate_up_proj_weight_quantizers.137 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.gate_up_proj_weight_quantizers.138 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.gate_up_proj_weight_quantizers.139 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.gate_up_proj_weight_quantizers.140 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.gate_up_proj_weight_quantizers.141 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.gate_up_proj_weight_quantizers.142 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.gate_up_proj_weight_quantizers.143 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.gate_up_proj_weight_quantizers.144 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.gate_up_proj_weight_quantizers.145 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.gate_up_proj_weight_quantizers.146 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.gate_up_proj_weight_quantizers.147 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.gate_up_proj_weight_quantizers.148 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.52e-01 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.gate_up_proj_weight_quantizers.149 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.88e-01 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.gate_up_proj_weight_quantizers.150 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.13e-01 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.gate_up_proj_weight_quantizers.151 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.gate_up_proj_weight_quantizers.152 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.95e-01 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.gate_up_proj_weight_quantizers.153 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.gate_up_proj_weight_quantizers.154 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.gate_up_proj_weight_quantizers.155 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.gate_up_proj_weight_quantizers.156 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.gate_up_proj_weight_quantizers.157 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.93e-01 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.gate_up_proj_weight_quantizers.158 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.gate_up_proj_weight_quantizers.159 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.gate_up_proj_weight_quantizers.160 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.gate_up_proj_weight_quantizers.161 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.gate_up_proj_weight_quantizers.162 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.75e-01 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.gate_up_proj_weight_quantizers.163 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.13e-01 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.gate_up_proj_weight_quantizers.164 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.gate_up_proj_weight_quantizers.165 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.27e-01 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.gate_up_proj_weight_quantizers.166 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.gate_up_proj_weight_quantizers.167 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.gate_up_proj_weight_quantizers.168 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.gate_up_proj_weight_quantizers.169 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.17e-01 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.gate_up_proj_weight_quantizers.170 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.gate_up_proj_weight_quantizers.171 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.gate_up_proj_weight_quantizers.172 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.32e-01 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.gate_up_proj_weight_quantizers.173 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.gate_up_proj_weight_quantizers.174 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.49e-01 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.gate_up_proj_weight_quantizers.175 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.gate_up_proj_weight_quantizers.176 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.40e-01 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.gate_up_proj_weight_quantizers.177 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.17e-01 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.gate_up_proj_weight_quantizers.178 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.gate_up_proj_weight_quantizers.179 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.gate_up_proj_weight_quantizers.180 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.51e-01 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.gate_up_proj_weight_quantizers.181 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.gate_up_proj_weight_quantizers.182 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.gate_up_proj_weight_quantizers.183 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.gate_up_proj_weight_quantizers.184 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.gate_up_proj_weight_quantizers.185 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.gate_up_proj_weight_quantizers.186 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.gate_up_proj_weight_quantizers.187 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.37e-01 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.gate_up_proj_weight_quantizers.188 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.gate_up_proj_weight_quantizers.189 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=4.94e-01 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.gate_up_proj_weight_quantizers.190 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.gate_up_proj_weight_quantizers.191 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.gate_up_proj_weight_quantizers.192 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.28e-01 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.gate_up_proj_weight_quantizers.193 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.gate_up_proj_weight_quantizers.194 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.gate_up_proj_weight_quantizers.195 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.gate_up_proj_weight_quantizers.196 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.gate_up_proj_weight_quantizers.197 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.gate_up_proj_weight_quantizers.198 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.gate_up_proj_weight_quantizers.199 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.gate_up_proj_weight_quantizers.200 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.13e-01 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.gate_up_proj_weight_quantizers.201 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.gate_up_proj_weight_quantizers.202 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.gate_up_proj_weight_quantizers.203 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.gate_up_proj_weight_quantizers.204 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.gate_up_proj_weight_quantizers.205 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.17e-01 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.gate_up_proj_weight_quantizers.206 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.gate_up_proj_weight_quantizers.207 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.gate_up_proj_weight_quantizers.208 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.gate_up_proj_weight_quantizers.209 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.gate_up_proj_weight_quantizers.210 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.gate_up_proj_weight_quantizers.211 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.gate_up_proj_weight_quantizers.212 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.gate_up_proj_weight_quantizers.213 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.gate_up_proj_weight_quantizers.214 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.15e-01 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.gate_up_proj_weight_quantizers.215 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.33e-01 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.gate_up_proj_weight_quantizers.216 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.gate_up_proj_weight_quantizers.217 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.gate_up_proj_weight_quantizers.218 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.gate_up_proj_weight_quantizers.219 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.gate_up_proj_weight_quantizers.220 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.14e-01 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.gate_up_proj_weight_quantizers.221 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.gate_up_proj_weight_quantizers.222 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.gate_up_proj_weight_quantizers.223 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.gate_up_proj_weight_quantizers.224 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.gate_up_proj_weight_quantizers.225 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=4.77e-01 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.gate_up_proj_weight_quantizers.226 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.gate_up_proj_weight_quantizers.227 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.gate_up_proj_weight_quantizers.228 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=5.43e-01 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.gate_up_proj_weight_quantizers.229 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.25e-01 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.gate_up_proj_weight_quantizers.230 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.gate_up_proj_weight_quantizers.231 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.gate_up_proj_weight_quantizers.232 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.gate_up_proj_weight_quantizers.233 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.gate_up_proj_weight_quantizers.234 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.gate_up_proj_weight_quantizers.235 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.gate_up_proj_weight_quantizers.236 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.71e-01 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.gate_up_proj_weight_quantizers.237 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.13e-01 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.gate_up_proj_weight_quantizers.238 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.14e-01 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.gate_up_proj_weight_quantizers.239 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.gate_up_proj_weight_quantizers.240 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.gate_up_proj_weight_quantizers.241 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.gate_up_proj_weight_quantizers.242 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.gate_up_proj_weight_quantizers.243 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.gate_up_proj_weight_quantizers.244 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.gate_up_proj_weight_quantizers.245 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.13e-01 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.gate_up_proj_weight_quantizers.246 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.gate_up_proj_weight_quantizers.247 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.gate_up_proj_weight_quantizers.248 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.gate_up_proj_weight_quantizers.249 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.gate_up_proj_weight_quantizers.250 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.16e-01 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.gate_up_proj_weight_quantizers.251 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.gate_up_proj_weight_quantizers.252 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.gate_up_proj_weight_quantizers.253 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.gate_up_proj_weight_quantizers.254 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.gate_up_proj_weight_quantizers.255 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.down_proj_input_quantizer TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.65e+01 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.down_proj_weight_quantizers.0 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.52e-01 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.down_proj_weight_quantizers.1 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.45e-01 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.down_proj_weight_quantizers.2 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.46e-01 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.down_proj_weight_quantizers.3 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.down_proj_weight_quantizers.4 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.down_proj_weight_quantizers.5 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.down_proj_weight_quantizers.6 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.down_proj_weight_quantizers.7 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.down_proj_weight_quantizers.8 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.59e-02 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.down_proj_weight_quantizers.9 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.down_proj_weight_quantizers.10 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.18e-01 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.down_proj_weight_quantizers.11 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.down_proj_weight_quantizers.12 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.down_proj_weight_quantizers.13 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.down_proj_weight_quantizers.14 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.down_proj_weight_quantizers.15 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.64e-02 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.down_proj_weight_quantizers.16 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.down_proj_weight_quantizers.17 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.down_proj_weight_quantizers.18 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.down_proj_weight_quantizers.19 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.14e-01 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.down_proj_weight_quantizers.20 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.down_proj_weight_quantizers.21 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.down_proj_weight_quantizers.22 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.down_proj_weight_quantizers.23 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.down_proj_weight_quantizers.24 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.down_proj_weight_quantizers.25 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.down_proj_weight_quantizers.26 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.down_proj_weight_quantizers.27 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.down_proj_weight_quantizers.28 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.down_proj_weight_quantizers.29 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.down_proj_weight_quantizers.30 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.32e-01 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.down_proj_weight_quantizers.31 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.down_proj_weight_quantizers.32 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.down_proj_weight_quantizers.33 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.down_proj_weight_quantizers.34 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.49e-01 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.down_proj_weight_quantizers.35 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.31e-01 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.down_proj_weight_quantizers.36 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.down_proj_weight_quantizers.37 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.down_proj_weight_quantizers.38 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.down_proj_weight_quantizers.39 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.down_proj_weight_quantizers.40 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.down_proj_weight_quantizers.41 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.down_proj_weight_quantizers.42 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.50e-01 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.down_proj_weight_quantizers.43 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.down_proj_weight_quantizers.44 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.down_proj_weight_quantizers.45 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.48e-01 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.down_proj_weight_quantizers.46 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.down_proj_weight_quantizers.47 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.down_proj_weight_quantizers.48 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.down_proj_weight_quantizers.49 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.97e-01 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.down_proj_weight_quantizers.50 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.down_proj_weight_quantizers.51 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.down_proj_weight_quantizers.52 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.down_proj_weight_quantizers.53 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.down_proj_weight_quantizers.54 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.down_proj_weight_quantizers.55 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.down_proj_weight_quantizers.56 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.down_proj_weight_quantizers.57 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.down_proj_weight_quantizers.58 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.56e-01 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.down_proj_weight_quantizers.59 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.down_proj_weight_quantizers.60 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.down_proj_weight_quantizers.61 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.down_proj_weight_quantizers.62 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.down_proj_weight_quantizers.63 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.70e-01 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.down_proj_weight_quantizers.64 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.down_proj_weight_quantizers.65 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.down_proj_weight_quantizers.66 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.54e-02 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.down_proj_weight_quantizers.67 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.down_proj_weight_quantizers.68 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.down_proj_weight_quantizers.69 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.down_proj_weight_quantizers.70 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.down_proj_weight_quantizers.71 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.down_proj_weight_quantizers.72 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.down_proj_weight_quantizers.73 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.down_proj_weight_quantizers.74 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.down_proj_weight_quantizers.75 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.down_proj_weight_quantizers.76 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.down_proj_weight_quantizers.77 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.down_proj_weight_quantizers.78 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.down_proj_weight_quantizers.79 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.down_proj_weight_quantizers.80 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.down_proj_weight_quantizers.81 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.down_proj_weight_quantizers.82 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.down_proj_weight_quantizers.83 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.64e-02 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.down_proj_weight_quantizers.84 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.down_proj_weight_quantizers.85 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.down_proj_weight_quantizers.86 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.down_proj_weight_quantizers.87 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.64e-02 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.down_proj_weight_quantizers.88 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.down_proj_weight_quantizers.89 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.down_proj_weight_quantizers.90 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.down_proj_weight_quantizers.91 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.down_proj_weight_quantizers.92 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.down_proj_weight_quantizers.93 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.down_proj_weight_quantizers.94 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.down_proj_weight_quantizers.95 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.down_proj_weight_quantizers.96 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.down_proj_weight_quantizers.97 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.down_proj_weight_quantizers.98 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.down_proj_weight_quantizers.99 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.down_proj_weight_quantizers.100 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.64e-02 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.down_proj_weight_quantizers.101 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.28e-01 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.down_proj_weight_quantizers.102 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.down_proj_weight_quantizers.103 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.down_proj_weight_quantizers.104 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.45e-01 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.down_proj_weight_quantizers.105 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.down_proj_weight_quantizers.106 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.down_proj_weight_quantizers.107 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.down_proj_weight_quantizers.108 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.down_proj_weight_quantizers.109 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.down_proj_weight_quantizers.110 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.down_proj_weight_quantizers.111 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.54e-02 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.down_proj_weight_quantizers.112 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.down_proj_weight_quantizers.113 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.down_proj_weight_quantizers.114 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.down_proj_weight_quantizers.115 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.down_proj_weight_quantizers.116 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.15e-01 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.down_proj_weight_quantizers.117 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.down_proj_weight_quantizers.118 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.down_proj_weight_quantizers.119 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.46e-01 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.down_proj_weight_quantizers.120 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.down_proj_weight_quantizers.121 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.down_proj_weight_quantizers.122 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.down_proj_weight_quantizers.123 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.down_proj_weight_quantizers.124 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.down_proj_weight_quantizers.125 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.down_proj_weight_quantizers.126 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.down_proj_weight_quantizers.127 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.down_proj_weight_quantizers.128 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.down_proj_weight_quantizers.129 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.30e-01 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.down_proj_weight_quantizers.130 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.down_proj_weight_quantizers.131 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.down_proj_weight_quantizers.132 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.23e-01 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.down_proj_weight_quantizers.133 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.down_proj_weight_quantizers.134 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.down_proj_weight_quantizers.135 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.down_proj_weight_quantizers.136 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.31e-01 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.down_proj_weight_quantizers.137 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.down_proj_weight_quantizers.138 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.down_proj_weight_quantizers.139 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.down_proj_weight_quantizers.140 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.down_proj_weight_quantizers.141 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.40e-01 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.down_proj_weight_quantizers.142 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.down_proj_weight_quantizers.143 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.down_proj_weight_quantizers.144 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.down_proj_weight_quantizers.145 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.down_proj_weight_quantizers.146 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.down_proj_weight_quantizers.147 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.down_proj_weight_quantizers.148 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.38e-01 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.down_proj_weight_quantizers.149 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.69e-01 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.down_proj_weight_quantizers.150 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.down_proj_weight_quantizers.151 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.down_proj_weight_quantizers.152 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.down_proj_weight_quantizers.153 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.down_proj_weight_quantizers.154 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.down_proj_weight_quantizers.155 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.down_proj_weight_quantizers.156 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.down_proj_weight_quantizers.157 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.down_proj_weight_quantizers.158 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.59e-02 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.down_proj_weight_quantizers.159 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.down_proj_weight_quantizers.160 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.down_proj_weight_quantizers.161 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.down_proj_weight_quantizers.162 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.62e-01 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.down_proj_weight_quantizers.163 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.20e-02 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.down_proj_weight_quantizers.164 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.down_proj_weight_quantizers.165 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.45e-02 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.down_proj_weight_quantizers.166 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.down_proj_weight_quantizers.167 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.down_proj_weight_quantizers.168 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.down_proj_weight_quantizers.169 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.down_proj_weight_quantizers.170 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.down_proj_weight_quantizers.171 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.67e-01 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.down_proj_weight_quantizers.172 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.down_proj_weight_quantizers.173 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.down_proj_weight_quantizers.174 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.29e-01 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.down_proj_weight_quantizers.175 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.down_proj_weight_quantizers.176 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.down_proj_weight_quantizers.177 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.down_proj_weight_quantizers.178 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.down_proj_weight_quantizers.179 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.64e-02 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.down_proj_weight_quantizers.180 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.27e-01 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.down_proj_weight_quantizers.181 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.down_proj_weight_quantizers.182 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.down_proj_weight_quantizers.183 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.down_proj_weight_quantizers.184 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.50e-02 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.down_proj_weight_quantizers.185 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.64e-02 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.down_proj_weight_quantizers.186 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.54e-02 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.down_proj_weight_quantizers.187 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.down_proj_weight_quantizers.188 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.down_proj_weight_quantizers.189 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.17e-01 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.down_proj_weight_quantizers.190 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.64e-02 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.down_proj_weight_quantizers.191 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.down_proj_weight_quantizers.192 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.down_proj_weight_quantizers.193 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.down_proj_weight_quantizers.194 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.down_proj_weight_quantizers.195 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.down_proj_weight_quantizers.196 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.down_proj_weight_quantizers.197 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.down_proj_weight_quantizers.198 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.down_proj_weight_quantizers.199 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.down_proj_weight_quantizers.200 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.down_proj_weight_quantizers.201 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.down_proj_weight_quantizers.202 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.down_proj_weight_quantizers.203 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.down_proj_weight_quantizers.204 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.down_proj_weight_quantizers.205 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.down_proj_weight_quantizers.206 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.down_proj_weight_quantizers.207 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.down_proj_weight_quantizers.208 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.down_proj_weight_quantizers.209 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.down_proj_weight_quantizers.210 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.down_proj_weight_quantizers.211 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.34e-01 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.down_proj_weight_quantizers.212 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.down_proj_weight_quantizers.213 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.38e-01 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.down_proj_weight_quantizers.214 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.70e-01 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.down_proj_weight_quantizers.215 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.17e-01 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.down_proj_weight_quantizers.216 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.down_proj_weight_quantizers.217 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.45e-02 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.down_proj_weight_quantizers.218 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.down_proj_weight_quantizers.219 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.down_proj_weight_quantizers.220 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.down_proj_weight_quantizers.221 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.down_proj_weight_quantizers.222 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.21e-01 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.down_proj_weight_quantizers.223 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.down_proj_weight_quantizers.224 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.down_proj_weight_quantizers.225 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.down_proj_weight_quantizers.226 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.down_proj_weight_quantizers.227 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.down_proj_weight_quantizers.228 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.down_proj_weight_quantizers.229 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.31e-01 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.down_proj_weight_quantizers.230 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.down_proj_weight_quantizers.231 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.13e-01 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.down_proj_weight_quantizers.232 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.down_proj_weight_quantizers.233 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.down_proj_weight_quantizers.234 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.down_proj_weight_quantizers.235 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.down_proj_weight_quantizers.236 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.63e-01 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.down_proj_weight_quantizers.237 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.down_proj_weight_quantizers.238 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.26e-01 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.down_proj_weight_quantizers.239 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.18e-01 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.down_proj_weight_quantizers.240 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.down_proj_weight_quantizers.241 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.42e-01 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.down_proj_weight_quantizers.242 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.down_proj_weight_quantizers.243 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.down_proj_weight_quantizers.244 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.61e-01 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.down_proj_weight_quantizers.245 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.down_proj_weight_quantizers.246 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.down_proj_weight_quantizers.247 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.down_proj_weight_quantizers.248 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.down_proj_weight_quantizers.249 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.down_proj_weight_quantizers.250 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.down_proj_weight_quantizers.251 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.down_proj_weight_quantizers.252 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.down_proj_weight_quantizers.253 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.down_proj_weight_quantizers.254 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.down_proj_weight_quantizers.255 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.61.mlp.shared_experts.gate_proj.input_quantizer TensorQuantizer(disabled) +model.layers.61.mlp.shared_experts.gate_proj.output_quantizer TensorQuantizer(disabled) +model.layers.61.mlp.shared_experts.gate_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.61.mlp.shared_experts.up_proj.input_quantizer TensorQuantizer(disabled) +model.layers.61.mlp.shared_experts.up_proj.output_quantizer TensorQuantizer(disabled) +model.layers.61.mlp.shared_experts.up_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.61.mlp.shared_experts.down_proj.input_quantizer TensorQuantizer(disabled) +model.layers.61.mlp.shared_experts.down_proj.output_quantizer TensorQuantizer(disabled) +model.layers.61.mlp.shared_experts.down_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.62.self_attn.q_a_proj.input_quantizer TensorQuantizer(disabled) +model.layers.62.self_attn.q_a_proj.output_quantizer TensorQuantizer(disabled) +model.layers.62.self_attn.q_a_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.62.self_attn.q_b_proj.input_quantizer TensorQuantizer(disabled) +model.layers.62.self_attn.q_b_proj.output_quantizer TensorQuantizer(disabled) +model.layers.62.self_attn.q_b_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.62.self_attn.kv_a_proj_with_mqa.input_quantizer TensorQuantizer(disabled) +model.layers.62.self_attn.kv_a_proj_with_mqa.output_quantizer TensorQuantizer(disabled) +model.layers.62.self_attn.kv_a_proj_with_mqa.weight_quantizer TensorQuantizer(disabled) +model.layers.62.self_attn.kv_b_proj.input_quantizer TensorQuantizer(disabled) +model.layers.62.self_attn.kv_b_proj.output_quantizer TensorQuantizer(disabled) +model.layers.62.self_attn.kv_b_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.62.self_attn.o_proj.input_quantizer TensorQuantizer(disabled) +model.layers.62.self_attn.o_proj.output_quantizer TensorQuantizer(disabled) +model.layers.62.self_attn.o_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.62.self_attn.indexer.wq_b.input_quantizer TensorQuantizer(disabled) +model.layers.62.self_attn.indexer.wq_b.output_quantizer TensorQuantizer(disabled) +model.layers.62.self_attn.indexer.wq_b.weight_quantizer TensorQuantizer(disabled) +model.layers.62.self_attn.indexer.wk.input_quantizer TensorQuantizer(disabled) +model.layers.62.self_attn.indexer.wk.output_quantizer TensorQuantizer(disabled) +model.layers.62.self_attn.indexer.wk.weight_quantizer TensorQuantizer(disabled) +model.layers.62.self_attn.indexer.k_norm.input_quantizer TensorQuantizer(disabled) +model.layers.62.self_attn.indexer.k_norm.output_quantizer TensorQuantizer(disabled) +model.layers.62.self_attn.indexer.weights_proj.input_quantizer TensorQuantizer(disabled) +model.layers.62.self_attn.indexer.weights_proj.output_quantizer TensorQuantizer(disabled) +model.layers.62.self_attn.indexer.weights_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.62.self_attn.q_bmm_quantizer TensorQuantizer(disabled) +model.layers.62.self_attn.k_bmm_quantizer TensorQuantizer((4, 3) bit fake per-tensor amax=4.48e+02(const) calibrator=MaxCalibrator quant) +model.layers.62.self_attn.v_bmm_quantizer TensorQuantizer((4, 3) bit fake per-tensor amax=4.48e+02(const) calibrator=MaxCalibrator quant) +model.layers.62.self_attn.p_bmm_quantizer TensorQuantizer(disabled) +model.layers.62.mlp.experts.gate_up_proj_input_quantizer TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.05e+00 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.gate_up_proj_weight_quantizers.0 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.96e-01 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.gate_up_proj_weight_quantizers.1 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.31e-01 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.gate_up_proj_weight_quantizers.2 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.15e-01 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.gate_up_proj_weight_quantizers.3 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.gate_up_proj_weight_quantizers.4 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.32e-01 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.gate_up_proj_weight_quantizers.5 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.gate_up_proj_weight_quantizers.6 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.21e-01 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.gate_up_proj_weight_quantizers.7 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.gate_up_proj_weight_quantizers.8 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.gate_up_proj_weight_quantizers.9 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.gate_up_proj_weight_quantizers.10 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.gate_up_proj_weight_quantizers.11 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.48e-01 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.gate_up_proj_weight_quantizers.12 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.gate_up_proj_weight_quantizers.13 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.13e-01 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.gate_up_proj_weight_quantizers.14 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.gate_up_proj_weight_quantizers.15 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.gate_up_proj_weight_quantizers.16 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.96e-01 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.gate_up_proj_weight_quantizers.17 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.gate_up_proj_weight_quantizers.18 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.gate_up_proj_weight_quantizers.19 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.gate_up_proj_weight_quantizers.20 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.gate_up_proj_weight_quantizers.21 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.gate_up_proj_weight_quantizers.22 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.gate_up_proj_weight_quantizers.23 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.gate_up_proj_weight_quantizers.24 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.gate_up_proj_weight_quantizers.25 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.gate_up_proj_weight_quantizers.26 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.gate_up_proj_weight_quantizers.27 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.gate_up_proj_weight_quantizers.28 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.gate_up_proj_weight_quantizers.29 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.17e-01 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.gate_up_proj_weight_quantizers.30 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.gate_up_proj_weight_quantizers.31 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.gate_up_proj_weight_quantizers.32 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.16e-01 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.gate_up_proj_weight_quantizers.33 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.16e-01 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.gate_up_proj_weight_quantizers.34 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.gate_up_proj_weight_quantizers.35 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.gate_up_proj_weight_quantizers.36 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.gate_up_proj_weight_quantizers.37 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.gate_up_proj_weight_quantizers.38 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.gate_up_proj_weight_quantizers.39 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.gate_up_proj_weight_quantizers.40 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.41e-01 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.gate_up_proj_weight_quantizers.41 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.gate_up_proj_weight_quantizers.42 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.gate_up_proj_weight_quantizers.43 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.89e-01 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.gate_up_proj_weight_quantizers.44 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.gate_up_proj_weight_quantizers.45 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.12e-01 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.gate_up_proj_weight_quantizers.46 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.gate_up_proj_weight_quantizers.47 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.gate_up_proj_weight_quantizers.48 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.gate_up_proj_weight_quantizers.49 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.28e-01 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.gate_up_proj_weight_quantizers.50 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.gate_up_proj_weight_quantizers.51 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.gate_up_proj_weight_quantizers.52 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.gate_up_proj_weight_quantizers.53 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.gate_up_proj_weight_quantizers.54 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.68e-01 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.gate_up_proj_weight_quantizers.55 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.gate_up_proj_weight_quantizers.56 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.gate_up_proj_weight_quantizers.57 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.gate_up_proj_weight_quantizers.58 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.gate_up_proj_weight_quantizers.59 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.gate_up_proj_weight_quantizers.60 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.gate_up_proj_weight_quantizers.61 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.gate_up_proj_weight_quantizers.62 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.gate_up_proj_weight_quantizers.63 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.19e-01 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.gate_up_proj_weight_quantizers.64 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.gate_up_proj_weight_quantizers.65 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.gate_up_proj_weight_quantizers.66 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.23e-01 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.gate_up_proj_weight_quantizers.67 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.gate_up_proj_weight_quantizers.68 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.gate_up_proj_weight_quantizers.69 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.17e-01 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.gate_up_proj_weight_quantizers.70 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.gate_up_proj_weight_quantizers.71 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.gate_up_proj_weight_quantizers.72 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.gate_up_proj_weight_quantizers.73 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.21e-01 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.gate_up_proj_weight_quantizers.74 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.gate_up_proj_weight_quantizers.75 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.gate_up_proj_weight_quantizers.76 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.19e-01 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.gate_up_proj_weight_quantizers.77 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.gate_up_proj_weight_quantizers.78 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.gate_up_proj_weight_quantizers.79 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.gate_up_proj_weight_quantizers.80 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.gate_up_proj_weight_quantizers.81 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.gate_up_proj_weight_quantizers.82 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.74e-01 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.gate_up_proj_weight_quantizers.83 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.gate_up_proj_weight_quantizers.84 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.gate_up_proj_weight_quantizers.85 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.gate_up_proj_weight_quantizers.86 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.gate_up_proj_weight_quantizers.87 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.gate_up_proj_weight_quantizers.88 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.30e-01 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.gate_up_proj_weight_quantizers.89 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.gate_up_proj_weight_quantizers.90 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.gate_up_proj_weight_quantizers.91 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.64e-02 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.gate_up_proj_weight_quantizers.92 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.gate_up_proj_weight_quantizers.93 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.gate_up_proj_weight_quantizers.94 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.gate_up_proj_weight_quantizers.95 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.gate_up_proj_weight_quantizers.96 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.gate_up_proj_weight_quantizers.97 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.gate_up_proj_weight_quantizers.98 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.gate_up_proj_weight_quantizers.99 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.46e-01 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.gate_up_proj_weight_quantizers.100 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.74e-01 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.gate_up_proj_weight_quantizers.101 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.85e-01 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.gate_up_proj_weight_quantizers.102 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.43e-01 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.gate_up_proj_weight_quantizers.103 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.gate_up_proj_weight_quantizers.104 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.gate_up_proj_weight_quantizers.105 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.gate_up_proj_weight_quantizers.106 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.gate_up_proj_weight_quantizers.107 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.89e-01 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.gate_up_proj_weight_quantizers.108 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.gate_up_proj_weight_quantizers.109 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.gate_up_proj_weight_quantizers.110 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.gate_up_proj_weight_quantizers.111 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.gate_up_proj_weight_quantizers.112 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.gate_up_proj_weight_quantizers.113 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.52e-01 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.gate_up_proj_weight_quantizers.114 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.gate_up_proj_weight_quantizers.115 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.gate_up_proj_weight_quantizers.116 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.76e-01 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.gate_up_proj_weight_quantizers.117 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.28e-01 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.gate_up_proj_weight_quantizers.118 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.gate_up_proj_weight_quantizers.119 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.81e-01 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.gate_up_proj_weight_quantizers.120 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.gate_up_proj_weight_quantizers.121 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.gate_up_proj_weight_quantizers.122 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.gate_up_proj_weight_quantizers.123 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.gate_up_proj_weight_quantizers.124 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.gate_up_proj_weight_quantizers.125 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.06e-01 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.gate_up_proj_weight_quantizers.126 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.gate_up_proj_weight_quantizers.127 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.46e-01 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.gate_up_proj_weight_quantizers.128 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.gate_up_proj_weight_quantizers.129 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.gate_up_proj_weight_quantizers.130 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.15e-01 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.gate_up_proj_weight_quantizers.131 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.gate_up_proj_weight_quantizers.132 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.gate_up_proj_weight_quantizers.133 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.gate_up_proj_weight_quantizers.134 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.gate_up_proj_weight_quantizers.135 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.gate_up_proj_weight_quantizers.136 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.gate_up_proj_weight_quantizers.137 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.gate_up_proj_weight_quantizers.138 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.28e-01 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.gate_up_proj_weight_quantizers.139 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.90e-01 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.gate_up_proj_weight_quantizers.140 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.gate_up_proj_weight_quantizers.141 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.gate_up_proj_weight_quantizers.142 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.41e-01 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.gate_up_proj_weight_quantizers.143 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.gate_up_proj_weight_quantizers.144 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.gate_up_proj_weight_quantizers.145 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.gate_up_proj_weight_quantizers.146 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.gate_up_proj_weight_quantizers.147 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.34e-01 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.gate_up_proj_weight_quantizers.148 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.39e-01 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.gate_up_proj_weight_quantizers.149 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.gate_up_proj_weight_quantizers.150 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.gate_up_proj_weight_quantizers.151 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.gate_up_proj_weight_quantizers.152 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.gate_up_proj_weight_quantizers.153 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.gate_up_proj_weight_quantizers.154 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.gate_up_proj_weight_quantizers.155 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.15e-01 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.gate_up_proj_weight_quantizers.156 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.gate_up_proj_weight_quantizers.157 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.gate_up_proj_weight_quantizers.158 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.13e-01 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.gate_up_proj_weight_quantizers.159 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.gate_up_proj_weight_quantizers.160 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.gate_up_proj_weight_quantizers.161 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.84e-01 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.gate_up_proj_weight_quantizers.162 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.81e-01 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.gate_up_proj_weight_quantizers.163 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.55e-01 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.gate_up_proj_weight_quantizers.164 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.48e-01 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.gate_up_proj_weight_quantizers.165 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.16e-01 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.gate_up_proj_weight_quantizers.166 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.gate_up_proj_weight_quantizers.167 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.17e-01 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.gate_up_proj_weight_quantizers.168 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.gate_up_proj_weight_quantizers.169 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.95e-01 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.gate_up_proj_weight_quantizers.170 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.gate_up_proj_weight_quantizers.171 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.gate_up_proj_weight_quantizers.172 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.gate_up_proj_weight_quantizers.173 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.31e-01 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.gate_up_proj_weight_quantizers.174 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.gate_up_proj_weight_quantizers.175 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.gate_up_proj_weight_quantizers.176 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.23e-01 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.gate_up_proj_weight_quantizers.177 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.gate_up_proj_weight_quantizers.178 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.gate_up_proj_weight_quantizers.179 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.gate_up_proj_weight_quantizers.180 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.gate_up_proj_weight_quantizers.181 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.gate_up_proj_weight_quantizers.182 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.gate_up_proj_weight_quantizers.183 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.77e-01 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.gate_up_proj_weight_quantizers.184 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.28e-01 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.gate_up_proj_weight_quantizers.185 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.gate_up_proj_weight_quantizers.186 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.05e-01 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.gate_up_proj_weight_quantizers.187 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.gate_up_proj_weight_quantizers.188 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.gate_up_proj_weight_quantizers.189 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.gate_up_proj_weight_quantizers.190 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.gate_up_proj_weight_quantizers.191 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.18e-01 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.gate_up_proj_weight_quantizers.192 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.56e-01 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.gate_up_proj_weight_quantizers.193 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.gate_up_proj_weight_quantizers.194 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.gate_up_proj_weight_quantizers.195 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.gate_up_proj_weight_quantizers.196 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.gate_up_proj_weight_quantizers.197 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.gate_up_proj_weight_quantizers.198 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.gate_up_proj_weight_quantizers.199 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.51e-01 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.gate_up_proj_weight_quantizers.200 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.45e-01 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.gate_up_proj_weight_quantizers.201 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.gate_up_proj_weight_quantizers.202 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.gate_up_proj_weight_quantizers.203 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.gate_up_proj_weight_quantizers.204 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.gate_up_proj_weight_quantizers.205 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.58e-01 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.gate_up_proj_weight_quantizers.206 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.gate_up_proj_weight_quantizers.207 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.79e-01 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.gate_up_proj_weight_quantizers.208 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.gate_up_proj_weight_quantizers.209 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.gate_up_proj_weight_quantizers.210 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.gate_up_proj_weight_quantizers.211 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.gate_up_proj_weight_quantizers.212 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.gate_up_proj_weight_quantizers.213 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.31e-01 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.gate_up_proj_weight_quantizers.214 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.gate_up_proj_weight_quantizers.215 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.gate_up_proj_weight_quantizers.216 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.59e-01 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.gate_up_proj_weight_quantizers.217 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.31e-01 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.gate_up_proj_weight_quantizers.218 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.gate_up_proj_weight_quantizers.219 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.gate_up_proj_weight_quantizers.220 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.gate_up_proj_weight_quantizers.221 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.gate_up_proj_weight_quantizers.222 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.gate_up_proj_weight_quantizers.223 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.gate_up_proj_weight_quantizers.224 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.gate_up_proj_weight_quantizers.225 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.gate_up_proj_weight_quantizers.226 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.gate_up_proj_weight_quantizers.227 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.gate_up_proj_weight_quantizers.228 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.gate_up_proj_weight_quantizers.229 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.gate_up_proj_weight_quantizers.230 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.gate_up_proj_weight_quantizers.231 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.gate_up_proj_weight_quantizers.232 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.gate_up_proj_weight_quantizers.233 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.gate_up_proj_weight_quantizers.234 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.gate_up_proj_weight_quantizers.235 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.gate_up_proj_weight_quantizers.236 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.gate_up_proj_weight_quantizers.237 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.gate_up_proj_weight_quantizers.238 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.24e-01 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.gate_up_proj_weight_quantizers.239 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.gate_up_proj_weight_quantizers.240 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.gate_up_proj_weight_quantizers.241 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.94e-01 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.gate_up_proj_weight_quantizers.242 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.gate_up_proj_weight_quantizers.243 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.gate_up_proj_weight_quantizers.244 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.gate_up_proj_weight_quantizers.245 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.52e-01 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.gate_up_proj_weight_quantizers.246 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.gate_up_proj_weight_quantizers.247 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.13e-01 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.gate_up_proj_weight_quantizers.248 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.gate_up_proj_weight_quantizers.249 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.gate_up_proj_weight_quantizers.250 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.gate_up_proj_weight_quantizers.251 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.46e-01 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.gate_up_proj_weight_quantizers.252 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.gate_up_proj_weight_quantizers.253 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.gate_up_proj_weight_quantizers.254 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.gate_up_proj_weight_quantizers.255 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.down_proj_input_quantizer TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=4.85e+01 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.down_proj_weight_quantizers.0 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.down_proj_weight_quantizers.1 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.down_proj_weight_quantizers.2 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.down_proj_weight_quantizers.3 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.down_proj_weight_quantizers.4 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.down_proj_weight_quantizers.5 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.down_proj_weight_quantizers.6 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.down_proj_weight_quantizers.7 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.down_proj_weight_quantizers.8 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.down_proj_weight_quantizers.9 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.down_proj_weight_quantizers.10 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.down_proj_weight_quantizers.11 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.down_proj_weight_quantizers.12 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.down_proj_weight_quantizers.13 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.down_proj_weight_quantizers.14 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.down_proj_weight_quantizers.15 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.45e-02 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.down_proj_weight_quantizers.16 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.down_proj_weight_quantizers.17 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.down_proj_weight_quantizers.18 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.down_proj_weight_quantizers.19 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.down_proj_weight_quantizers.20 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.down_proj_weight_quantizers.21 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.76e-01 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.down_proj_weight_quantizers.22 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.down_proj_weight_quantizers.23 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.down_proj_weight_quantizers.24 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.down_proj_weight_quantizers.25 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.down_proj_weight_quantizers.26 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.down_proj_weight_quantizers.27 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.down_proj_weight_quantizers.28 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.down_proj_weight_quantizers.29 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.down_proj_weight_quantizers.30 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.54e-02 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.down_proj_weight_quantizers.31 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.11e-02 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.down_proj_weight_quantizers.32 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.down_proj_weight_quantizers.33 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.down_proj_weight_quantizers.34 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.down_proj_weight_quantizers.35 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.down_proj_weight_quantizers.36 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.down_proj_weight_quantizers.37 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.down_proj_weight_quantizers.38 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.down_proj_weight_quantizers.39 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.64e-02 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.down_proj_weight_quantizers.40 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.down_proj_weight_quantizers.41 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.down_proj_weight_quantizers.42 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.down_proj_weight_quantizers.43 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.down_proj_weight_quantizers.44 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.82e-01 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.down_proj_weight_quantizers.45 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.45e-01 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.down_proj_weight_quantizers.46 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.down_proj_weight_quantizers.47 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.54e-02 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.down_proj_weight_quantizers.48 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.down_proj_weight_quantizers.49 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.31e-01 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.down_proj_weight_quantizers.50 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.67e-01 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.down_proj_weight_quantizers.51 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.down_proj_weight_quantizers.52 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.down_proj_weight_quantizers.53 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.down_proj_weight_quantizers.54 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=5.35e-01 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.down_proj_weight_quantizers.55 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.down_proj_weight_quantizers.56 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.down_proj_weight_quantizers.57 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.down_proj_weight_quantizers.58 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.down_proj_weight_quantizers.59 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.42e-01 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.down_proj_weight_quantizers.60 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.down_proj_weight_quantizers.61 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.down_proj_weight_quantizers.62 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.down_proj_weight_quantizers.63 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.down_proj_weight_quantizers.64 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.down_proj_weight_quantizers.65 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.down_proj_weight_quantizers.66 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.down_proj_weight_quantizers.67 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.down_proj_weight_quantizers.68 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.11e-01 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.down_proj_weight_quantizers.69 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.down_proj_weight_quantizers.70 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.down_proj_weight_quantizers.71 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.down_proj_weight_quantizers.72 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.down_proj_weight_quantizers.73 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.down_proj_weight_quantizers.74 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.down_proj_weight_quantizers.75 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.down_proj_weight_quantizers.76 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.down_proj_weight_quantizers.77 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.down_proj_weight_quantizers.78 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.down_proj_weight_quantizers.79 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.down_proj_weight_quantizers.80 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.down_proj_weight_quantizers.81 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.down_proj_weight_quantizers.82 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.37e-01 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.down_proj_weight_quantizers.83 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.down_proj_weight_quantizers.84 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.15e-01 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.down_proj_weight_quantizers.85 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.down_proj_weight_quantizers.86 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.down_proj_weight_quantizers.87 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.down_proj_weight_quantizers.88 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.down_proj_weight_quantizers.89 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.down_proj_weight_quantizers.90 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.down_proj_weight_quantizers.91 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.down_proj_weight_quantizers.92 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.64e-02 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.down_proj_weight_quantizers.93 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.down_proj_weight_quantizers.94 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.down_proj_weight_quantizers.95 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.down_proj_weight_quantizers.96 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.down_proj_weight_quantizers.97 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.down_proj_weight_quantizers.98 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.down_proj_weight_quantizers.99 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.16e-01 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.down_proj_weight_quantizers.100 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.22e-01 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.down_proj_weight_quantizers.101 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.down_proj_weight_quantizers.102 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.down_proj_weight_quantizers.103 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.down_proj_weight_quantizers.104 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.down_proj_weight_quantizers.105 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.down_proj_weight_quantizers.106 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.64e-02 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.down_proj_weight_quantizers.107 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.68e-01 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.down_proj_weight_quantizers.108 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.down_proj_weight_quantizers.109 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.down_proj_weight_quantizers.110 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.down_proj_weight_quantizers.111 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.down_proj_weight_quantizers.112 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.down_proj_weight_quantizers.113 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.down_proj_weight_quantizers.114 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.down_proj_weight_quantizers.115 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.down_proj_weight_quantizers.116 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.down_proj_weight_quantizers.117 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.54e-02 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.down_proj_weight_quantizers.118 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.down_proj_weight_quantizers.119 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.48e-01 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.down_proj_weight_quantizers.120 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.down_proj_weight_quantizers.121 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.down_proj_weight_quantizers.122 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.23e-01 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.down_proj_weight_quantizers.123 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.down_proj_weight_quantizers.124 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.down_proj_weight_quantizers.125 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.60e-01 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.down_proj_weight_quantizers.126 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.down_proj_weight_quantizers.127 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.down_proj_weight_quantizers.128 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.down_proj_weight_quantizers.129 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.down_proj_weight_quantizers.130 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.down_proj_weight_quantizers.131 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.93e-01 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.down_proj_weight_quantizers.132 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.down_proj_weight_quantizers.133 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.down_proj_weight_quantizers.134 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.down_proj_weight_quantizers.135 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.29e-01 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.down_proj_weight_quantizers.136 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.down_proj_weight_quantizers.137 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.down_proj_weight_quantizers.138 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.39e-01 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.down_proj_weight_quantizers.139 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.down_proj_weight_quantizers.140 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.down_proj_weight_quantizers.141 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.down_proj_weight_quantizers.142 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.down_proj_weight_quantizers.143 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.down_proj_weight_quantizers.144 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.down_proj_weight_quantizers.145 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.down_proj_weight_quantizers.146 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.18e-01 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.down_proj_weight_quantizers.147 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.down_proj_weight_quantizers.148 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.down_proj_weight_quantizers.149 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.down_proj_weight_quantizers.150 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.down_proj_weight_quantizers.151 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.down_proj_weight_quantizers.152 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.down_proj_weight_quantizers.153 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.down_proj_weight_quantizers.154 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.down_proj_weight_quantizers.155 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.down_proj_weight_quantizers.156 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.down_proj_weight_quantizers.157 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.down_proj_weight_quantizers.158 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.down_proj_weight_quantizers.159 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.down_proj_weight_quantizers.160 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.down_proj_weight_quantizers.161 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.down_proj_weight_quantizers.162 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.67e-01 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.down_proj_weight_quantizers.163 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.down_proj_weight_quantizers.164 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.down_proj_weight_quantizers.165 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.down_proj_weight_quantizers.166 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.down_proj_weight_quantizers.167 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.down_proj_weight_quantizers.168 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.15e-01 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.down_proj_weight_quantizers.169 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.down_proj_weight_quantizers.170 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.down_proj_weight_quantizers.171 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.down_proj_weight_quantizers.172 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.59e-02 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.down_proj_weight_quantizers.173 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.down_proj_weight_quantizers.174 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.down_proj_weight_quantizers.175 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.down_proj_weight_quantizers.176 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.down_proj_weight_quantizers.177 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.down_proj_weight_quantizers.178 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.down_proj_weight_quantizers.179 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.down_proj_weight_quantizers.180 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.down_proj_weight_quantizers.181 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.down_proj_weight_quantizers.182 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.down_proj_weight_quantizers.183 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.86e-01 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.down_proj_weight_quantizers.184 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.80e-01 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.down_proj_weight_quantizers.185 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.down_proj_weight_quantizers.186 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.46e-01 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.down_proj_weight_quantizers.187 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.down_proj_weight_quantizers.188 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.down_proj_weight_quantizers.189 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.down_proj_weight_quantizers.190 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.down_proj_weight_quantizers.191 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.down_proj_weight_quantizers.192 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.60e-01 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.down_proj_weight_quantizers.193 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.down_proj_weight_quantizers.194 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.down_proj_weight_quantizers.195 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.down_proj_weight_quantizers.196 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.down_proj_weight_quantizers.197 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.down_proj_weight_quantizers.198 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.14e-01 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.down_proj_weight_quantizers.199 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.down_proj_weight_quantizers.200 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.down_proj_weight_quantizers.201 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.down_proj_weight_quantizers.202 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.down_proj_weight_quantizers.203 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.down_proj_weight_quantizers.204 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.down_proj_weight_quantizers.205 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.down_proj_weight_quantizers.206 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.59e-02 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.down_proj_weight_quantizers.207 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=5.35e-01 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.down_proj_weight_quantizers.208 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.down_proj_weight_quantizers.209 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.down_proj_weight_quantizers.210 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.15e-01 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.down_proj_weight_quantizers.211 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.down_proj_weight_quantizers.212 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.down_proj_weight_quantizers.213 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.64e-02 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.down_proj_weight_quantizers.214 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.down_proj_weight_quantizers.215 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.59e-02 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.down_proj_weight_quantizers.216 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.50e-01 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.down_proj_weight_quantizers.217 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.24e-01 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.down_proj_weight_quantizers.218 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.down_proj_weight_quantizers.219 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.down_proj_weight_quantizers.220 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.down_proj_weight_quantizers.221 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.down_proj_weight_quantizers.222 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.down_proj_weight_quantizers.223 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.down_proj_weight_quantizers.224 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.down_proj_weight_quantizers.225 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.down_proj_weight_quantizers.226 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.54e-02 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.down_proj_weight_quantizers.227 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.down_proj_weight_quantizers.228 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.35e-02 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.down_proj_weight_quantizers.229 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.down_proj_weight_quantizers.230 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.down_proj_weight_quantizers.231 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.50e-02 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.down_proj_weight_quantizers.232 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.down_proj_weight_quantizers.233 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.down_proj_weight_quantizers.234 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.21e-01 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.down_proj_weight_quantizers.235 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.down_proj_weight_quantizers.236 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.down_proj_weight_quantizers.237 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.down_proj_weight_quantizers.238 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.27e-01 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.down_proj_weight_quantizers.239 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.down_proj_weight_quantizers.240 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.down_proj_weight_quantizers.241 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.82e-01 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.down_proj_weight_quantizers.242 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.down_proj_weight_quantizers.243 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.down_proj_weight_quantizers.244 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.down_proj_weight_quantizers.245 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.34e-01 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.down_proj_weight_quantizers.246 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.down_proj_weight_quantizers.247 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.down_proj_weight_quantizers.248 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.45e-01 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.down_proj_weight_quantizers.249 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.down_proj_weight_quantizers.250 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.down_proj_weight_quantizers.251 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.down_proj_weight_quantizers.252 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.down_proj_weight_quantizers.253 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.down_proj_weight_quantizers.254 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.down_proj_weight_quantizers.255 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.62.mlp.shared_experts.gate_proj.input_quantizer TensorQuantizer(disabled) +model.layers.62.mlp.shared_experts.gate_proj.output_quantizer TensorQuantizer(disabled) +model.layers.62.mlp.shared_experts.gate_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.62.mlp.shared_experts.up_proj.input_quantizer TensorQuantizer(disabled) +model.layers.62.mlp.shared_experts.up_proj.output_quantizer TensorQuantizer(disabled) +model.layers.62.mlp.shared_experts.up_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.62.mlp.shared_experts.down_proj.input_quantizer TensorQuantizer(disabled) +model.layers.62.mlp.shared_experts.down_proj.output_quantizer TensorQuantizer(disabled) +model.layers.62.mlp.shared_experts.down_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.63.self_attn.q_a_proj.input_quantizer TensorQuantizer(disabled) +model.layers.63.self_attn.q_a_proj.output_quantizer TensorQuantizer(disabled) +model.layers.63.self_attn.q_a_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.63.self_attn.q_b_proj.input_quantizer TensorQuantizer(disabled) +model.layers.63.self_attn.q_b_proj.output_quantizer TensorQuantizer(disabled) +model.layers.63.self_attn.q_b_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.63.self_attn.kv_a_proj_with_mqa.input_quantizer TensorQuantizer(disabled) +model.layers.63.self_attn.kv_a_proj_with_mqa.output_quantizer TensorQuantizer(disabled) +model.layers.63.self_attn.kv_a_proj_with_mqa.weight_quantizer TensorQuantizer(disabled) +model.layers.63.self_attn.kv_b_proj.input_quantizer TensorQuantizer(disabled) +model.layers.63.self_attn.kv_b_proj.output_quantizer TensorQuantizer(disabled) +model.layers.63.self_attn.kv_b_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.63.self_attn.o_proj.input_quantizer TensorQuantizer(disabled) +model.layers.63.self_attn.o_proj.output_quantizer TensorQuantizer(disabled) +model.layers.63.self_attn.o_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.63.self_attn.q_bmm_quantizer TensorQuantizer(disabled) +model.layers.63.self_attn.k_bmm_quantizer TensorQuantizer((4, 3) bit fake per-tensor amax=4.48e+02(const) calibrator=MaxCalibrator quant) +model.layers.63.self_attn.v_bmm_quantizer TensorQuantizer((4, 3) bit fake per-tensor amax=4.48e+02(const) calibrator=MaxCalibrator quant) +model.layers.63.self_attn.p_bmm_quantizer TensorQuantizer(disabled) +model.layers.63.mlp.experts.gate_up_proj_input_quantizer TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.92e+00 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.gate_up_proj_weight_quantizers.0 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.gate_up_proj_weight_quantizers.1 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.gate_up_proj_weight_quantizers.2 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.gate_up_proj_weight_quantizers.3 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.27e-01 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.gate_up_proj_weight_quantizers.4 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.gate_up_proj_weight_quantizers.5 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.55e-01 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.gate_up_proj_weight_quantizers.6 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.62e-01 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.gate_up_proj_weight_quantizers.7 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.gate_up_proj_weight_quantizers.8 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.gate_up_proj_weight_quantizers.9 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.gate_up_proj_weight_quantizers.10 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.gate_up_proj_weight_quantizers.11 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.gate_up_proj_weight_quantizers.12 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.gate_up_proj_weight_quantizers.13 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.gate_up_proj_weight_quantizers.14 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.34e-01 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.gate_up_proj_weight_quantizers.15 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.gate_up_proj_weight_quantizers.16 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.39e-01 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.gate_up_proj_weight_quantizers.17 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.gate_up_proj_weight_quantizers.18 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.97e-01 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.gate_up_proj_weight_quantizers.19 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.gate_up_proj_weight_quantizers.20 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.gate_up_proj_weight_quantizers.21 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.gate_up_proj_weight_quantizers.22 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.gate_up_proj_weight_quantizers.23 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.gate_up_proj_weight_quantizers.24 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.gate_up_proj_weight_quantizers.25 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.gate_up_proj_weight_quantizers.26 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.gate_up_proj_weight_quantizers.27 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.gate_up_proj_weight_quantizers.28 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.gate_up_proj_weight_quantizers.29 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.gate_up_proj_weight_quantizers.30 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.gate_up_proj_weight_quantizers.31 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.gate_up_proj_weight_quantizers.32 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.gate_up_proj_weight_quantizers.33 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.66e-01 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.gate_up_proj_weight_quantizers.34 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=4.59e-01 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.gate_up_proj_weight_quantizers.35 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.gate_up_proj_weight_quantizers.36 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.gate_up_proj_weight_quantizers.37 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.gate_up_proj_weight_quantizers.38 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.gate_up_proj_weight_quantizers.39 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.26e-01 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.gate_up_proj_weight_quantizers.40 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.gate_up_proj_weight_quantizers.41 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.gate_up_proj_weight_quantizers.42 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.gate_up_proj_weight_quantizers.43 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.gate_up_proj_weight_quantizers.44 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.88e-01 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.gate_up_proj_weight_quantizers.45 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.16e-01 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.gate_up_proj_weight_quantizers.46 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.gate_up_proj_weight_quantizers.47 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.gate_up_proj_weight_quantizers.48 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.41e-01 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.gate_up_proj_weight_quantizers.49 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.gate_up_proj_weight_quantizers.50 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.94e+00 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.gate_up_proj_weight_quantizers.51 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.gate_up_proj_weight_quantizers.52 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.79e-01 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.gate_up_proj_weight_quantizers.53 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.gate_up_proj_weight_quantizers.54 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.gate_up_proj_weight_quantizers.55 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.gate_up_proj_weight_quantizers.56 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.gate_up_proj_weight_quantizers.57 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.gate_up_proj_weight_quantizers.58 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.gate_up_proj_weight_quantizers.59 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.gate_up_proj_weight_quantizers.60 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.95e-01 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.gate_up_proj_weight_quantizers.61 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.gate_up_proj_weight_quantizers.62 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.gate_up_proj_weight_quantizers.63 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.gate_up_proj_weight_quantizers.64 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.gate_up_proj_weight_quantizers.65 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.gate_up_proj_weight_quantizers.66 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.gate_up_proj_weight_quantizers.67 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.gate_up_proj_weight_quantizers.68 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.29e-01 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.gate_up_proj_weight_quantizers.69 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.gate_up_proj_weight_quantizers.70 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.gate_up_proj_weight_quantizers.71 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.33e-01 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.gate_up_proj_weight_quantizers.72 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.28e-01 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.gate_up_proj_weight_quantizers.73 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.gate_up_proj_weight_quantizers.74 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.13e-01 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.gate_up_proj_weight_quantizers.75 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.gate_up_proj_weight_quantizers.76 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.gate_up_proj_weight_quantizers.77 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.gate_up_proj_weight_quantizers.78 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.49e-01 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.gate_up_proj_weight_quantizers.79 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.gate_up_proj_weight_quantizers.80 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.66e-01 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.gate_up_proj_weight_quantizers.81 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.gate_up_proj_weight_quantizers.82 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.gate_up_proj_weight_quantizers.83 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.gate_up_proj_weight_quantizers.84 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.gate_up_proj_weight_quantizers.85 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.gate_up_proj_weight_quantizers.86 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.gate_up_proj_weight_quantizers.87 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.gate_up_proj_weight_quantizers.88 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.gate_up_proj_weight_quantizers.89 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.gate_up_proj_weight_quantizers.90 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.gate_up_proj_weight_quantizers.91 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.gate_up_proj_weight_quantizers.92 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.14e-01 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.gate_up_proj_weight_quantizers.93 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.gate_up_proj_weight_quantizers.94 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.gate_up_proj_weight_quantizers.95 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.26e-01 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.gate_up_proj_weight_quantizers.96 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.46e-01 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.gate_up_proj_weight_quantizers.97 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.gate_up_proj_weight_quantizers.98 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.gate_up_proj_weight_quantizers.99 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.gate_up_proj_weight_quantizers.100 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.gate_up_proj_weight_quantizers.101 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.gate_up_proj_weight_quantizers.102 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.gate_up_proj_weight_quantizers.103 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.gate_up_proj_weight_quantizers.104 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.gate_up_proj_weight_quantizers.105 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.gate_up_proj_weight_quantizers.106 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.17e-01 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.gate_up_proj_weight_quantizers.107 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.39e-01 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.gate_up_proj_weight_quantizers.108 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.25e-01 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.gate_up_proj_weight_quantizers.109 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.gate_up_proj_weight_quantizers.110 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.gate_up_proj_weight_quantizers.111 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.gate_up_proj_weight_quantizers.112 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.gate_up_proj_weight_quantizers.113 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.25e-01 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.gate_up_proj_weight_quantizers.114 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.13e-01 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.gate_up_proj_weight_quantizers.115 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.gate_up_proj_weight_quantizers.116 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.33e-01 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.gate_up_proj_weight_quantizers.117 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.gate_up_proj_weight_quantizers.118 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.gate_up_proj_weight_quantizers.119 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.gate_up_proj_weight_quantizers.120 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.24e-01 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.gate_up_proj_weight_quantizers.121 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.gate_up_proj_weight_quantizers.122 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.61e-01 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.gate_up_proj_weight_quantizers.123 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.gate_up_proj_weight_quantizers.124 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.17e-01 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.gate_up_proj_weight_quantizers.125 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.gate_up_proj_weight_quantizers.126 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.61e-01 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.gate_up_proj_weight_quantizers.127 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.gate_up_proj_weight_quantizers.128 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.gate_up_proj_weight_quantizers.129 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.gate_up_proj_weight_quantizers.130 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.gate_up_proj_weight_quantizers.131 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.gate_up_proj_weight_quantizers.132 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.gate_up_proj_weight_quantizers.133 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.gate_up_proj_weight_quantizers.134 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.42e-01 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.gate_up_proj_weight_quantizers.135 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.gate_up_proj_weight_quantizers.136 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.52e-01 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.gate_up_proj_weight_quantizers.137 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.gate_up_proj_weight_quantizers.138 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.gate_up_proj_weight_quantizers.139 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.68e-01 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.gate_up_proj_weight_quantizers.140 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.gate_up_proj_weight_quantizers.141 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.gate_up_proj_weight_quantizers.142 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.gate_up_proj_weight_quantizers.143 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.63e-01 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.gate_up_proj_weight_quantizers.144 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.gate_up_proj_weight_quantizers.145 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.gate_up_proj_weight_quantizers.146 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.81e-01 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.gate_up_proj_weight_quantizers.147 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.gate_up_proj_weight_quantizers.148 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.35e-01 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.gate_up_proj_weight_quantizers.149 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.59e-02 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.gate_up_proj_weight_quantizers.150 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.gate_up_proj_weight_quantizers.151 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.gate_up_proj_weight_quantizers.152 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.72e-01 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.gate_up_proj_weight_quantizers.153 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.gate_up_proj_weight_quantizers.154 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.gate_up_proj_weight_quantizers.155 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.gate_up_proj_weight_quantizers.156 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.94e-01 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.gate_up_proj_weight_quantizers.157 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.gate_up_proj_weight_quantizers.158 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.gate_up_proj_weight_quantizers.159 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.gate_up_proj_weight_quantizers.160 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.gate_up_proj_weight_quantizers.161 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.24e-01 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.gate_up_proj_weight_quantizers.162 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.gate_up_proj_weight_quantizers.163 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.54e-01 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.gate_up_proj_weight_quantizers.164 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.gate_up_proj_weight_quantizers.165 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.gate_up_proj_weight_quantizers.166 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.gate_up_proj_weight_quantizers.167 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.gate_up_proj_weight_quantizers.168 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.gate_up_proj_weight_quantizers.169 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.gate_up_proj_weight_quantizers.170 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.gate_up_proj_weight_quantizers.171 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.gate_up_proj_weight_quantizers.172 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.gate_up_proj_weight_quantizers.173 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.62e-01 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.gate_up_proj_weight_quantizers.174 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.gate_up_proj_weight_quantizers.175 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.59e-01 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.gate_up_proj_weight_quantizers.176 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.46e-01 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.gate_up_proj_weight_quantizers.177 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.gate_up_proj_weight_quantizers.178 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.25e-01 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.gate_up_proj_weight_quantizers.179 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.gate_up_proj_weight_quantizers.180 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.gate_up_proj_weight_quantizers.181 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.71e-01 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.gate_up_proj_weight_quantizers.182 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.gate_up_proj_weight_quantizers.183 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.14e-01 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.gate_up_proj_weight_quantizers.184 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.gate_up_proj_weight_quantizers.185 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.gate_up_proj_weight_quantizers.186 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.gate_up_proj_weight_quantizers.187 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.65e-01 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.gate_up_proj_weight_quantizers.188 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.gate_up_proj_weight_quantizers.189 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.gate_up_proj_weight_quantizers.190 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.gate_up_proj_weight_quantizers.191 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.gate_up_proj_weight_quantizers.192 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.gate_up_proj_weight_quantizers.193 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.gate_up_proj_weight_quantizers.194 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.45e-01 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.gate_up_proj_weight_quantizers.195 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.65e-01 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.gate_up_proj_weight_quantizers.196 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.gate_up_proj_weight_quantizers.197 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.24e-01 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.gate_up_proj_weight_quantizers.198 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.gate_up_proj_weight_quantizers.199 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.59e-01 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.gate_up_proj_weight_quantizers.200 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.gate_up_proj_weight_quantizers.201 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.gate_up_proj_weight_quantizers.202 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.gate_up_proj_weight_quantizers.203 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.gate_up_proj_weight_quantizers.204 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.14e-01 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.gate_up_proj_weight_quantizers.205 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.gate_up_proj_weight_quantizers.206 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.gate_up_proj_weight_quantizers.207 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.gate_up_proj_weight_quantizers.208 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.gate_up_proj_weight_quantizers.209 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.gate_up_proj_weight_quantizers.210 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.gate_up_proj_weight_quantizers.211 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.14e-01 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.gate_up_proj_weight_quantizers.212 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.gate_up_proj_weight_quantizers.213 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.gate_up_proj_weight_quantizers.214 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.gate_up_proj_weight_quantizers.215 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.gate_up_proj_weight_quantizers.216 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.84e-01 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.gate_up_proj_weight_quantizers.217 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.gate_up_proj_weight_quantizers.218 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.gate_up_proj_weight_quantizers.219 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.gate_up_proj_weight_quantizers.220 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.gate_up_proj_weight_quantizers.221 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.gate_up_proj_weight_quantizers.222 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.gate_up_proj_weight_quantizers.223 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.21e-01 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.gate_up_proj_weight_quantizers.224 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.gate_up_proj_weight_quantizers.225 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.19e-01 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.gate_up_proj_weight_quantizers.226 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.69e-01 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.gate_up_proj_weight_quantizers.227 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.24e-01 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.gate_up_proj_weight_quantizers.228 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.28e-01 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.gate_up_proj_weight_quantizers.229 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.92e-01 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.gate_up_proj_weight_quantizers.230 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.gate_up_proj_weight_quantizers.231 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.gate_up_proj_weight_quantizers.232 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.gate_up_proj_weight_quantizers.233 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.59e-01 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.gate_up_proj_weight_quantizers.234 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.gate_up_proj_weight_quantizers.235 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.gate_up_proj_weight_quantizers.236 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.gate_up_proj_weight_quantizers.237 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.gate_up_proj_weight_quantizers.238 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=4.20e-01 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.gate_up_proj_weight_quantizers.239 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.54e-01 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.gate_up_proj_weight_quantizers.240 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.65e-01 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.gate_up_proj_weight_quantizers.241 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.gate_up_proj_weight_quantizers.242 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.gate_up_proj_weight_quantizers.243 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.gate_up_proj_weight_quantizers.244 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.gate_up_proj_weight_quantizers.245 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.gate_up_proj_weight_quantizers.246 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.14e-01 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.gate_up_proj_weight_quantizers.247 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.gate_up_proj_weight_quantizers.248 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.gate_up_proj_weight_quantizers.249 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.gate_up_proj_weight_quantizers.250 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.gate_up_proj_weight_quantizers.251 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.gate_up_proj_weight_quantizers.252 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.gate_up_proj_weight_quantizers.253 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.gate_up_proj_weight_quantizers.254 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.37e-01 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.gate_up_proj_weight_quantizers.255 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.15e-01 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.down_proj_input_quantizer TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=4.42e+01 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.down_proj_weight_quantizers.0 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.down_proj_weight_quantizers.1 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.down_proj_weight_quantizers.2 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.down_proj_weight_quantizers.3 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.down_proj_weight_quantizers.4 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.down_proj_weight_quantizers.5 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.42e-01 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.down_proj_weight_quantizers.6 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.down_proj_weight_quantizers.7 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.down_proj_weight_quantizers.8 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.down_proj_weight_quantizers.9 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.down_proj_weight_quantizers.10 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.down_proj_weight_quantizers.11 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.down_proj_weight_quantizers.12 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.down_proj_weight_quantizers.13 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.down_proj_weight_quantizers.14 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.down_proj_weight_quantizers.15 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.down_proj_weight_quantizers.16 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.15e-01 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.down_proj_weight_quantizers.17 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.down_proj_weight_quantizers.18 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.36e-01 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.down_proj_weight_quantizers.19 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.down_proj_weight_quantizers.20 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.down_proj_weight_quantizers.21 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.59e-02 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.down_proj_weight_quantizers.22 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.down_proj_weight_quantizers.23 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.14e-01 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.down_proj_weight_quantizers.24 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.down_proj_weight_quantizers.25 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.down_proj_weight_quantizers.26 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.down_proj_weight_quantizers.27 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.down_proj_weight_quantizers.28 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.down_proj_weight_quantizers.29 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.down_proj_weight_quantizers.30 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.down_proj_weight_quantizers.31 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.down_proj_weight_quantizers.32 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.down_proj_weight_quantizers.33 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.32e-01 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.down_proj_weight_quantizers.34 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.down_proj_weight_quantizers.35 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.down_proj_weight_quantizers.36 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.down_proj_weight_quantizers.37 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.50e-02 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.down_proj_weight_quantizers.38 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.down_proj_weight_quantizers.39 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.down_proj_weight_quantizers.40 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.54e-02 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.down_proj_weight_quantizers.41 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.down_proj_weight_quantizers.42 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.down_proj_weight_quantizers.43 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.down_proj_weight_quantizers.44 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.27e-01 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.down_proj_weight_quantizers.45 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.down_proj_weight_quantizers.46 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.down_proj_weight_quantizers.47 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.32e-01 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.down_proj_weight_quantizers.48 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.down_proj_weight_quantizers.49 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.down_proj_weight_quantizers.50 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.83e-01 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.down_proj_weight_quantizers.51 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.down_proj_weight_quantizers.52 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.73e-01 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.down_proj_weight_quantizers.53 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.down_proj_weight_quantizers.54 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.down_proj_weight_quantizers.55 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.21e-01 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.down_proj_weight_quantizers.56 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.down_proj_weight_quantizers.57 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.down_proj_weight_quantizers.58 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.down_proj_weight_quantizers.59 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.down_proj_weight_quantizers.60 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.51e-01 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.down_proj_weight_quantizers.61 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.down_proj_weight_quantizers.62 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.down_proj_weight_quantizers.63 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.down_proj_weight_quantizers.64 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.down_proj_weight_quantizers.65 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.down_proj_weight_quantizers.66 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.down_proj_weight_quantizers.67 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.down_proj_weight_quantizers.68 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.52e-01 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.down_proj_weight_quantizers.69 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.down_proj_weight_quantizers.70 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.down_proj_weight_quantizers.71 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.down_proj_weight_quantizers.72 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.down_proj_weight_quantizers.73 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.down_proj_weight_quantizers.74 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.down_proj_weight_quantizers.75 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.down_proj_weight_quantizers.76 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.down_proj_weight_quantizers.77 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.down_proj_weight_quantizers.78 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.down_proj_weight_quantizers.79 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.down_proj_weight_quantizers.80 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.86e-01 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.down_proj_weight_quantizers.81 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.down_proj_weight_quantizers.82 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.21e-01 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.down_proj_weight_quantizers.83 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.down_proj_weight_quantizers.84 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.down_proj_weight_quantizers.85 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.down_proj_weight_quantizers.86 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.down_proj_weight_quantizers.87 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.down_proj_weight_quantizers.88 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.down_proj_weight_quantizers.89 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.down_proj_weight_quantizers.90 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.down_proj_weight_quantizers.91 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.down_proj_weight_quantizers.92 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.down_proj_weight_quantizers.93 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.down_proj_weight_quantizers.94 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.down_proj_weight_quantizers.95 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.down_proj_weight_quantizers.96 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.15e-01 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.down_proj_weight_quantizers.97 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.down_proj_weight_quantizers.98 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.down_proj_weight_quantizers.99 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.45e-01 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.down_proj_weight_quantizers.100 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.down_proj_weight_quantizers.101 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.down_proj_weight_quantizers.102 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.down_proj_weight_quantizers.103 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.down_proj_weight_quantizers.104 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.27e-01 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.down_proj_weight_quantizers.105 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.down_proj_weight_quantizers.106 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.40e-01 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.down_proj_weight_quantizers.107 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.down_proj_weight_quantizers.108 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.23e-01 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.down_proj_weight_quantizers.109 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.down_proj_weight_quantizers.110 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.down_proj_weight_quantizers.111 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.down_proj_weight_quantizers.112 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.down_proj_weight_quantizers.113 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.down_proj_weight_quantizers.114 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.down_proj_weight_quantizers.115 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.down_proj_weight_quantizers.116 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.75e-01 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.down_proj_weight_quantizers.117 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.down_proj_weight_quantizers.118 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.down_proj_weight_quantizers.119 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.53e-01 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.down_proj_weight_quantizers.120 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.down_proj_weight_quantizers.121 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.down_proj_weight_quantizers.122 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.down_proj_weight_quantizers.123 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.down_proj_weight_quantizers.124 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.down_proj_weight_quantizers.125 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.down_proj_weight_quantizers.126 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.down_proj_weight_quantizers.127 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.75e-01 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.down_proj_weight_quantizers.128 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.down_proj_weight_quantizers.129 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.27e-01 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.down_proj_weight_quantizers.130 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.down_proj_weight_quantizers.131 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.down_proj_weight_quantizers.132 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.down_proj_weight_quantizers.133 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.down_proj_weight_quantizers.134 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.down_proj_weight_quantizers.135 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.down_proj_weight_quantizers.136 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.11e-01 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.down_proj_weight_quantizers.137 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.down_proj_weight_quantizers.138 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.down_proj_weight_quantizers.139 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.00e-01 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.down_proj_weight_quantizers.140 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.down_proj_weight_quantizers.141 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.down_proj_weight_quantizers.142 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.down_proj_weight_quantizers.143 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.89e-01 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.down_proj_weight_quantizers.144 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.down_proj_weight_quantizers.145 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.down_proj_weight_quantizers.146 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.down_proj_weight_quantizers.147 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.down_proj_weight_quantizers.148 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.47e-01 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.down_proj_weight_quantizers.149 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.down_proj_weight_quantizers.150 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.down_proj_weight_quantizers.151 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.down_proj_weight_quantizers.152 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.30e-01 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.down_proj_weight_quantizers.153 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.down_proj_weight_quantizers.154 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.down_proj_weight_quantizers.155 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.down_proj_weight_quantizers.156 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.down_proj_weight_quantizers.157 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.down_proj_weight_quantizers.158 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.down_proj_weight_quantizers.159 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.down_proj_weight_quantizers.160 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.down_proj_weight_quantizers.161 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.40e-01 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.down_proj_weight_quantizers.162 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.down_proj_weight_quantizers.163 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.55e-01 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.down_proj_weight_quantizers.164 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.54e-02 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.down_proj_weight_quantizers.165 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.down_proj_weight_quantizers.166 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.down_proj_weight_quantizers.167 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.down_proj_weight_quantizers.168 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.down_proj_weight_quantizers.169 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.down_proj_weight_quantizers.170 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.93e-01 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.down_proj_weight_quantizers.171 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.down_proj_weight_quantizers.172 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.down_proj_weight_quantizers.173 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.down_proj_weight_quantizers.174 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.down_proj_weight_quantizers.175 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.27e-01 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.down_proj_weight_quantizers.176 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.down_proj_weight_quantizers.177 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.down_proj_weight_quantizers.178 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.63e-01 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.down_proj_weight_quantizers.179 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.down_proj_weight_quantizers.180 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.54e-02 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.down_proj_weight_quantizers.181 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.00e-01 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.down_proj_weight_quantizers.182 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.down_proj_weight_quantizers.183 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.down_proj_weight_quantizers.184 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.down_proj_weight_quantizers.185 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.down_proj_weight_quantizers.186 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.down_proj_weight_quantizers.187 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.down_proj_weight_quantizers.188 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.down_proj_weight_quantizers.189 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.down_proj_weight_quantizers.190 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.down_proj_weight_quantizers.191 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.down_proj_weight_quantizers.192 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.down_proj_weight_quantizers.193 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.down_proj_weight_quantizers.194 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.down_proj_weight_quantizers.195 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.down_proj_weight_quantizers.196 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.down_proj_weight_quantizers.197 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.down_proj_weight_quantizers.198 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.30e-01 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.down_proj_weight_quantizers.199 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.down_proj_weight_quantizers.200 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.down_proj_weight_quantizers.201 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.down_proj_weight_quantizers.202 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.down_proj_weight_quantizers.203 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.down_proj_weight_quantizers.204 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.30e-01 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.down_proj_weight_quantizers.205 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.down_proj_weight_quantizers.206 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.down_proj_weight_quantizers.207 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.16e-01 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.down_proj_weight_quantizers.208 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.64e-02 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.down_proj_weight_quantizers.209 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.down_proj_weight_quantizers.210 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.31e-01 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.down_proj_weight_quantizers.211 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.85e-01 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.down_proj_weight_quantizers.212 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.down_proj_weight_quantizers.213 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.down_proj_weight_quantizers.214 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.down_proj_weight_quantizers.215 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.down_proj_weight_quantizers.216 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.48e-01 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.down_proj_weight_quantizers.217 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.down_proj_weight_quantizers.218 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.down_proj_weight_quantizers.219 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.down_proj_weight_quantizers.220 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.down_proj_weight_quantizers.221 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.down_proj_weight_quantizers.222 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.down_proj_weight_quantizers.223 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.45e-01 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.down_proj_weight_quantizers.224 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.64e-02 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.down_proj_weight_quantizers.225 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.down_proj_weight_quantizers.226 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.40e-01 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.down_proj_weight_quantizers.227 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.down_proj_weight_quantizers.228 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.down_proj_weight_quantizers.229 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.down_proj_weight_quantizers.230 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.down_proj_weight_quantizers.231 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.down_proj_weight_quantizers.232 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.down_proj_weight_quantizers.233 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.34e-01 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.down_proj_weight_quantizers.234 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.58e-01 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.down_proj_weight_quantizers.235 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.down_proj_weight_quantizers.236 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.down_proj_weight_quantizers.237 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.down_proj_weight_quantizers.238 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.37e-01 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.down_proj_weight_quantizers.239 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.50e-01 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.down_proj_weight_quantizers.240 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.38e-01 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.down_proj_weight_quantizers.241 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.down_proj_weight_quantizers.242 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.down_proj_weight_quantizers.243 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.down_proj_weight_quantizers.244 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.down_proj_weight_quantizers.245 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.down_proj_weight_quantizers.246 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.down_proj_weight_quantizers.247 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.down_proj_weight_quantizers.248 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.down_proj_weight_quantizers.249 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.down_proj_weight_quantizers.250 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.down_proj_weight_quantizers.251 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.down_proj_weight_quantizers.252 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.down_proj_weight_quantizers.253 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.down_proj_weight_quantizers.254 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.59e-02 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.down_proj_weight_quantizers.255 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.63.mlp.shared_experts.gate_proj.input_quantizer TensorQuantizer(disabled) +model.layers.63.mlp.shared_experts.gate_proj.output_quantizer TensorQuantizer(disabled) +model.layers.63.mlp.shared_experts.gate_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.63.mlp.shared_experts.up_proj.input_quantizer TensorQuantizer(disabled) +model.layers.63.mlp.shared_experts.up_proj.output_quantizer TensorQuantizer(disabled) +model.layers.63.mlp.shared_experts.up_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.63.mlp.shared_experts.down_proj.input_quantizer TensorQuantizer(disabled) +model.layers.63.mlp.shared_experts.down_proj.output_quantizer TensorQuantizer(disabled) +model.layers.63.mlp.shared_experts.down_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.64.self_attn.q_a_proj.input_quantizer TensorQuantizer(disabled) +model.layers.64.self_attn.q_a_proj.output_quantizer TensorQuantizer(disabled) +model.layers.64.self_attn.q_a_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.64.self_attn.q_b_proj.input_quantizer TensorQuantizer(disabled) +model.layers.64.self_attn.q_b_proj.output_quantizer TensorQuantizer(disabled) +model.layers.64.self_attn.q_b_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.64.self_attn.kv_a_proj_with_mqa.input_quantizer TensorQuantizer(disabled) +model.layers.64.self_attn.kv_a_proj_with_mqa.output_quantizer TensorQuantizer(disabled) +model.layers.64.self_attn.kv_a_proj_with_mqa.weight_quantizer TensorQuantizer(disabled) +model.layers.64.self_attn.kv_b_proj.input_quantizer TensorQuantizer(disabled) +model.layers.64.self_attn.kv_b_proj.output_quantizer TensorQuantizer(disabled) +model.layers.64.self_attn.kv_b_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.64.self_attn.o_proj.input_quantizer TensorQuantizer(disabled) +model.layers.64.self_attn.o_proj.output_quantizer TensorQuantizer(disabled) +model.layers.64.self_attn.o_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.64.self_attn.q_bmm_quantizer TensorQuantizer(disabled) +model.layers.64.self_attn.k_bmm_quantizer TensorQuantizer((4, 3) bit fake per-tensor amax=4.48e+02(const) calibrator=MaxCalibrator quant) +model.layers.64.self_attn.v_bmm_quantizer TensorQuantizer((4, 3) bit fake per-tensor amax=4.48e+02(const) calibrator=MaxCalibrator quant) +model.layers.64.self_attn.p_bmm_quantizer TensorQuantizer(disabled) +model.layers.64.mlp.experts.gate_up_proj_input_quantizer TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.97e+00 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.gate_up_proj_weight_quantizers.0 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.gate_up_proj_weight_quantizers.1 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.17e-01 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.gate_up_proj_weight_quantizers.2 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.38e-01 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.gate_up_proj_weight_quantizers.3 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.gate_up_proj_weight_quantizers.4 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.gate_up_proj_weight_quantizers.5 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.gate_up_proj_weight_quantizers.6 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.gate_up_proj_weight_quantizers.7 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.37e-01 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.gate_up_proj_weight_quantizers.8 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.gate_up_proj_weight_quantizers.9 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.57e-01 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.gate_up_proj_weight_quantizers.10 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.gate_up_proj_weight_quantizers.11 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.gate_up_proj_weight_quantizers.12 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.gate_up_proj_weight_quantizers.13 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.gate_up_proj_weight_quantizers.14 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.gate_up_proj_weight_quantizers.15 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.gate_up_proj_weight_quantizers.16 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.gate_up_proj_weight_quantizers.17 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.gate_up_proj_weight_quantizers.18 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.26e-01 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.gate_up_proj_weight_quantizers.19 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.gate_up_proj_weight_quantizers.20 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.92e-01 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.gate_up_proj_weight_quantizers.21 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.gate_up_proj_weight_quantizers.22 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.16e-01 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.gate_up_proj_weight_quantizers.23 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.gate_up_proj_weight_quantizers.24 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.gate_up_proj_weight_quantizers.25 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.46e-01 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.gate_up_proj_weight_quantizers.26 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.60e-01 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.gate_up_proj_weight_quantizers.27 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.gate_up_proj_weight_quantizers.28 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.47e-01 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.gate_up_proj_weight_quantizers.29 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.15e-01 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.gate_up_proj_weight_quantizers.30 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.13e-01 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.gate_up_proj_weight_quantizers.31 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.gate_up_proj_weight_quantizers.32 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.gate_up_proj_weight_quantizers.33 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.gate_up_proj_weight_quantizers.34 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.29e-01 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.gate_up_proj_weight_quantizers.35 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.gate_up_proj_weight_quantizers.36 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.gate_up_proj_weight_quantizers.37 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.gate_up_proj_weight_quantizers.38 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.gate_up_proj_weight_quantizers.39 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.20e-01 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.gate_up_proj_weight_quantizers.40 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.gate_up_proj_weight_quantizers.41 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.gate_up_proj_weight_quantizers.42 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.20e-01 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.gate_up_proj_weight_quantizers.43 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.gate_up_proj_weight_quantizers.44 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.26e-01 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.gate_up_proj_weight_quantizers.45 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.gate_up_proj_weight_quantizers.46 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.gate_up_proj_weight_quantizers.47 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.gate_up_proj_weight_quantizers.48 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.gate_up_proj_weight_quantizers.49 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.20e-01 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.gate_up_proj_weight_quantizers.50 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.gate_up_proj_weight_quantizers.51 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.32e-01 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.gate_up_proj_weight_quantizers.52 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.gate_up_proj_weight_quantizers.53 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.18e-01 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.gate_up_proj_weight_quantizers.54 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.41e-01 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.gate_up_proj_weight_quantizers.55 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.25e-01 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.gate_up_proj_weight_quantizers.56 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.gate_up_proj_weight_quantizers.57 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.gate_up_proj_weight_quantizers.58 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.21e-01 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.gate_up_proj_weight_quantizers.59 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.23e-01 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.gate_up_proj_weight_quantizers.60 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.gate_up_proj_weight_quantizers.61 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.20e-01 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.gate_up_proj_weight_quantizers.62 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.gate_up_proj_weight_quantizers.63 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.29e-01 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.gate_up_proj_weight_quantizers.64 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.gate_up_proj_weight_quantizers.65 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.gate_up_proj_weight_quantizers.66 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.gate_up_proj_weight_quantizers.67 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.gate_up_proj_weight_quantizers.68 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.gate_up_proj_weight_quantizers.69 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.25e-01 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.gate_up_proj_weight_quantizers.70 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.14e-01 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.gate_up_proj_weight_quantizers.71 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.72e-01 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.gate_up_proj_weight_quantizers.72 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.gate_up_proj_weight_quantizers.73 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.55e-01 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.gate_up_proj_weight_quantizers.74 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.gate_up_proj_weight_quantizers.75 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.gate_up_proj_weight_quantizers.76 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.gate_up_proj_weight_quantizers.77 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.gate_up_proj_weight_quantizers.78 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.gate_up_proj_weight_quantizers.79 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.gate_up_proj_weight_quantizers.80 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.16e-01 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.gate_up_proj_weight_quantizers.81 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.68e-01 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.gate_up_proj_weight_quantizers.82 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.gate_up_proj_weight_quantizers.83 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.gate_up_proj_weight_quantizers.84 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.gate_up_proj_weight_quantizers.85 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.gate_up_proj_weight_quantizers.86 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.gate_up_proj_weight_quantizers.87 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.gate_up_proj_weight_quantizers.88 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.gate_up_proj_weight_quantizers.89 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.gate_up_proj_weight_quantizers.90 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.gate_up_proj_weight_quantizers.91 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.gate_up_proj_weight_quantizers.92 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.gate_up_proj_weight_quantizers.93 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.gate_up_proj_weight_quantizers.94 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.62e-01 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.gate_up_proj_weight_quantizers.95 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.68e-01 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.gate_up_proj_weight_quantizers.96 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.24e-01 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.gate_up_proj_weight_quantizers.97 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.24e-01 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.gate_up_proj_weight_quantizers.98 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.14e-01 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.gate_up_proj_weight_quantizers.99 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.gate_up_proj_weight_quantizers.100 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=5.82e-01 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.gate_up_proj_weight_quantizers.101 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.gate_up_proj_weight_quantizers.102 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.gate_up_proj_weight_quantizers.103 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.gate_up_proj_weight_quantizers.104 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.gate_up_proj_weight_quantizers.105 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.gate_up_proj_weight_quantizers.106 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.gate_up_proj_weight_quantizers.107 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.gate_up_proj_weight_quantizers.108 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.gate_up_proj_weight_quantizers.109 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.gate_up_proj_weight_quantizers.110 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.52e-01 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.gate_up_proj_weight_quantizers.111 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.gate_up_proj_weight_quantizers.112 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.13e-01 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.gate_up_proj_weight_quantizers.113 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.gate_up_proj_weight_quantizers.114 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.15e-01 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.gate_up_proj_weight_quantizers.115 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.gate_up_proj_weight_quantizers.116 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.44e-01 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.gate_up_proj_weight_quantizers.117 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.gate_up_proj_weight_quantizers.118 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.gate_up_proj_weight_quantizers.119 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.gate_up_proj_weight_quantizers.120 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.gate_up_proj_weight_quantizers.121 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.gate_up_proj_weight_quantizers.122 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.gate_up_proj_weight_quantizers.123 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.91e-01 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.gate_up_proj_weight_quantizers.124 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.28e-01 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.gate_up_proj_weight_quantizers.125 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.21e-01 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.gate_up_proj_weight_quantizers.126 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.36e-01 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.gate_up_proj_weight_quantizers.127 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.17e-01 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.gate_up_proj_weight_quantizers.128 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.gate_up_proj_weight_quantizers.129 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.gate_up_proj_weight_quantizers.130 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.gate_up_proj_weight_quantizers.131 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.gate_up_proj_weight_quantizers.132 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.gate_up_proj_weight_quantizers.133 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.gate_up_proj_weight_quantizers.134 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.gate_up_proj_weight_quantizers.135 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.gate_up_proj_weight_quantizers.136 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.19e-01 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.gate_up_proj_weight_quantizers.137 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.77e-01 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.gate_up_proj_weight_quantizers.138 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.gate_up_proj_weight_quantizers.139 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.gate_up_proj_weight_quantizers.140 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.gate_up_proj_weight_quantizers.141 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.gate_up_proj_weight_quantizers.142 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.gate_up_proj_weight_quantizers.143 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.57e-01 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.gate_up_proj_weight_quantizers.144 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.gate_up_proj_weight_quantizers.145 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.gate_up_proj_weight_quantizers.146 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.23e-01 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.gate_up_proj_weight_quantizers.147 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.gate_up_proj_weight_quantizers.148 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.gate_up_proj_weight_quantizers.149 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.gate_up_proj_weight_quantizers.150 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.gate_up_proj_weight_quantizers.151 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.15e-01 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.gate_up_proj_weight_quantizers.152 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.26e-01 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.gate_up_proj_weight_quantizers.153 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.gate_up_proj_weight_quantizers.154 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.gate_up_proj_weight_quantizers.155 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.gate_up_proj_weight_quantizers.156 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.gate_up_proj_weight_quantizers.157 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.73e-01 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.gate_up_proj_weight_quantizers.158 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.gate_up_proj_weight_quantizers.159 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.gate_up_proj_weight_quantizers.160 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.37e-01 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.gate_up_proj_weight_quantizers.161 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.gate_up_proj_weight_quantizers.162 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.17e-01 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.gate_up_proj_weight_quantizers.163 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.15e-01 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.gate_up_proj_weight_quantizers.164 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.gate_up_proj_weight_quantizers.165 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.gate_up_proj_weight_quantizers.166 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.gate_up_proj_weight_quantizers.167 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.gate_up_proj_weight_quantizers.168 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.42e-01 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.gate_up_proj_weight_quantizers.169 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.gate_up_proj_weight_quantizers.170 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.46e-01 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.gate_up_proj_weight_quantizers.171 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.gate_up_proj_weight_quantizers.172 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.gate_up_proj_weight_quantizers.173 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.gate_up_proj_weight_quantizers.174 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.gate_up_proj_weight_quantizers.175 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.17e-01 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.gate_up_proj_weight_quantizers.176 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.16e-01 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.gate_up_proj_weight_quantizers.177 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.gate_up_proj_weight_quantizers.178 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.gate_up_proj_weight_quantizers.179 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.gate_up_proj_weight_quantizers.180 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.88e-01 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.gate_up_proj_weight_quantizers.181 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.gate_up_proj_weight_quantizers.182 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.gate_up_proj_weight_quantizers.183 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=4.06e-01 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.gate_up_proj_weight_quantizers.184 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.gate_up_proj_weight_quantizers.185 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.17e-01 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.gate_up_proj_weight_quantizers.186 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.41e-01 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.gate_up_proj_weight_quantizers.187 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.gate_up_proj_weight_quantizers.188 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.gate_up_proj_weight_quantizers.189 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.gate_up_proj_weight_quantizers.190 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.75e-01 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.gate_up_proj_weight_quantizers.191 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.gate_up_proj_weight_quantizers.192 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.gate_up_proj_weight_quantizers.193 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.gate_up_proj_weight_quantizers.194 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.gate_up_proj_weight_quantizers.195 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.25e-01 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.gate_up_proj_weight_quantizers.196 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.gate_up_proj_weight_quantizers.197 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.gate_up_proj_weight_quantizers.198 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.gate_up_proj_weight_quantizers.199 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.gate_up_proj_weight_quantizers.200 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.gate_up_proj_weight_quantizers.201 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.14e-01 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.gate_up_proj_weight_quantizers.202 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.14e-01 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.gate_up_proj_weight_quantizers.203 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.gate_up_proj_weight_quantizers.204 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.62e-01 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.gate_up_proj_weight_quantizers.205 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.13e-01 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.gate_up_proj_weight_quantizers.206 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.gate_up_proj_weight_quantizers.207 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.gate_up_proj_weight_quantizers.208 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.gate_up_proj_weight_quantizers.209 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.gate_up_proj_weight_quantizers.210 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.gate_up_proj_weight_quantizers.211 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.gate_up_proj_weight_quantizers.212 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.gate_up_proj_weight_quantizers.213 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.19e-01 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.gate_up_proj_weight_quantizers.214 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.gate_up_proj_weight_quantizers.215 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.71e-01 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.gate_up_proj_weight_quantizers.216 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.gate_up_proj_weight_quantizers.217 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.36e-01 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.gate_up_proj_weight_quantizers.218 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.gate_up_proj_weight_quantizers.219 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.54e-01 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.gate_up_proj_weight_quantizers.220 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.gate_up_proj_weight_quantizers.221 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.gate_up_proj_weight_quantizers.222 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.gate_up_proj_weight_quantizers.223 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.14e-01 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.gate_up_proj_weight_quantizers.224 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.gate_up_proj_weight_quantizers.225 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.gate_up_proj_weight_quantizers.226 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.gate_up_proj_weight_quantizers.227 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.gate_up_proj_weight_quantizers.228 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.gate_up_proj_weight_quantizers.229 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.gate_up_proj_weight_quantizers.230 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.55e-01 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.gate_up_proj_weight_quantizers.231 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.gate_up_proj_weight_quantizers.232 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.gate_up_proj_weight_quantizers.233 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.13e-01 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.gate_up_proj_weight_quantizers.234 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.gate_up_proj_weight_quantizers.235 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.gate_up_proj_weight_quantizers.236 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.gate_up_proj_weight_quantizers.237 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.gate_up_proj_weight_quantizers.238 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.gate_up_proj_weight_quantizers.239 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.58e-01 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.gate_up_proj_weight_quantizers.240 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.20e-01 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.gate_up_proj_weight_quantizers.241 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.gate_up_proj_weight_quantizers.242 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.gate_up_proj_weight_quantizers.243 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.gate_up_proj_weight_quantizers.244 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.gate_up_proj_weight_quantizers.245 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.17e-01 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.gate_up_proj_weight_quantizers.246 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.gate_up_proj_weight_quantizers.247 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.gate_up_proj_weight_quantizers.248 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.gate_up_proj_weight_quantizers.249 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.gate_up_proj_weight_quantizers.250 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.37e-01 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.gate_up_proj_weight_quantizers.251 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.gate_up_proj_weight_quantizers.252 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.14e-01 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.gate_up_proj_weight_quantizers.253 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.58e-01 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.gate_up_proj_weight_quantizers.254 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.gate_up_proj_weight_quantizers.255 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.16e-01 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.down_proj_input_quantizer TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=5.75e+01 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.down_proj_weight_quantizers.0 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.64e-02 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.down_proj_weight_quantizers.1 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.down_proj_weight_quantizers.2 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.down_proj_weight_quantizers.3 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.down_proj_weight_quantizers.4 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.down_proj_weight_quantizers.5 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.down_proj_weight_quantizers.6 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.down_proj_weight_quantizers.7 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.down_proj_weight_quantizers.8 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.down_proj_weight_quantizers.9 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.down_proj_weight_quantizers.10 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.down_proj_weight_quantizers.11 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.down_proj_weight_quantizers.12 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.down_proj_weight_quantizers.13 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.down_proj_weight_quantizers.14 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.down_proj_weight_quantizers.15 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.29e-01 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.down_proj_weight_quantizers.16 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.down_proj_weight_quantizers.17 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.64e-02 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.down_proj_weight_quantizers.18 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.17e-01 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.down_proj_weight_quantizers.19 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.down_proj_weight_quantizers.20 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.79e-01 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.down_proj_weight_quantizers.21 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.down_proj_weight_quantizers.22 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.20e-01 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.down_proj_weight_quantizers.23 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.40e-01 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.down_proj_weight_quantizers.24 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.04e-01 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.down_proj_weight_quantizers.25 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.down_proj_weight_quantizers.26 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.down_proj_weight_quantizers.27 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.66e-01 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.down_proj_weight_quantizers.28 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.79e-01 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.down_proj_weight_quantizers.29 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.down_proj_weight_quantizers.30 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.down_proj_weight_quantizers.31 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.down_proj_weight_quantizers.32 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.down_proj_weight_quantizers.33 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.down_proj_weight_quantizers.34 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.down_proj_weight_quantizers.35 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.down_proj_weight_quantizers.36 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.down_proj_weight_quantizers.37 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.down_proj_weight_quantizers.38 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.down_proj_weight_quantizers.39 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.down_proj_weight_quantizers.40 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.down_proj_weight_quantizers.41 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.down_proj_weight_quantizers.42 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.down_proj_weight_quantizers.43 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.down_proj_weight_quantizers.44 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.20e-01 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.down_proj_weight_quantizers.45 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.80e-01 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.down_proj_weight_quantizers.46 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.down_proj_weight_quantizers.47 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.down_proj_weight_quantizers.48 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.down_proj_weight_quantizers.49 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.down_proj_weight_quantizers.50 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.down_proj_weight_quantizers.51 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.down_proj_weight_quantizers.52 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.down_proj_weight_quantizers.53 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.70e-01 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.down_proj_weight_quantizers.54 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.70e-01 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.down_proj_weight_quantizers.55 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.down_proj_weight_quantizers.56 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.down_proj_weight_quantizers.57 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.down_proj_weight_quantizers.58 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.down_proj_weight_quantizers.59 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.down_proj_weight_quantizers.60 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.down_proj_weight_quantizers.61 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.67e-01 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.down_proj_weight_quantizers.62 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.down_proj_weight_quantizers.63 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.down_proj_weight_quantizers.64 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.down_proj_weight_quantizers.65 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.down_proj_weight_quantizers.66 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.down_proj_weight_quantizers.67 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.down_proj_weight_quantizers.68 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.35e-02 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.down_proj_weight_quantizers.69 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.down_proj_weight_quantizers.70 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.down_proj_weight_quantizers.71 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.30e-01 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.down_proj_weight_quantizers.72 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.down_proj_weight_quantizers.73 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.down_proj_weight_quantizers.74 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.50e-02 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.down_proj_weight_quantizers.75 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.down_proj_weight_quantizers.76 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.down_proj_weight_quantizers.77 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.down_proj_weight_quantizers.78 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.down_proj_weight_quantizers.79 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.down_proj_weight_quantizers.80 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.26e-01 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.down_proj_weight_quantizers.81 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.down_proj_weight_quantizers.82 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.down_proj_weight_quantizers.83 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.down_proj_weight_quantizers.84 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.down_proj_weight_quantizers.85 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.down_proj_weight_quantizers.86 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.down_proj_weight_quantizers.87 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.28e-01 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.down_proj_weight_quantizers.88 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.down_proj_weight_quantizers.89 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.down_proj_weight_quantizers.90 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.down_proj_weight_quantizers.91 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.down_proj_weight_quantizers.92 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.down_proj_weight_quantizers.93 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.down_proj_weight_quantizers.94 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.38e-01 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.down_proj_weight_quantizers.95 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.23e-01 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.down_proj_weight_quantizers.96 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.down_proj_weight_quantizers.97 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.down_proj_weight_quantizers.98 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.54e-02 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.down_proj_weight_quantizers.99 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.71e-01 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.down_proj_weight_quantizers.100 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.69e-01 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.down_proj_weight_quantizers.101 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.down_proj_weight_quantizers.102 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.down_proj_weight_quantizers.103 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.down_proj_weight_quantizers.104 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.down_proj_weight_quantizers.105 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.down_proj_weight_quantizers.106 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.down_proj_weight_quantizers.107 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.down_proj_weight_quantizers.108 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.down_proj_weight_quantizers.109 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.12e-01 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.down_proj_weight_quantizers.110 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.down_proj_weight_quantizers.111 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.down_proj_weight_quantizers.112 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.down_proj_weight_quantizers.113 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.down_proj_weight_quantizers.114 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.59e-02 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.down_proj_weight_quantizers.115 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.down_proj_weight_quantizers.116 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.down_proj_weight_quantizers.117 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.down_proj_weight_quantizers.118 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.down_proj_weight_quantizers.119 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.down_proj_weight_quantizers.120 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.down_proj_weight_quantizers.121 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.down_proj_weight_quantizers.122 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.down_proj_weight_quantizers.123 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.31e-01 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.down_proj_weight_quantizers.124 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.36e-01 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.down_proj_weight_quantizers.125 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.down_proj_weight_quantizers.126 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.19e-01 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.down_proj_weight_quantizers.127 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.down_proj_weight_quantizers.128 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.down_proj_weight_quantizers.129 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.down_proj_weight_quantizers.130 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.down_proj_weight_quantizers.131 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.down_proj_weight_quantizers.132 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.down_proj_weight_quantizers.133 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.64e-02 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.down_proj_weight_quantizers.134 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.down_proj_weight_quantizers.135 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.down_proj_weight_quantizers.136 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.down_proj_weight_quantizers.137 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.down_proj_weight_quantizers.138 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.down_proj_weight_quantizers.139 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.down_proj_weight_quantizers.140 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.down_proj_weight_quantizers.141 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.down_proj_weight_quantizers.142 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.down_proj_weight_quantizers.143 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.28e-01 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.down_proj_weight_quantizers.144 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.down_proj_weight_quantizers.145 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.down_proj_weight_quantizers.146 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.down_proj_weight_quantizers.147 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.down_proj_weight_quantizers.148 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.down_proj_weight_quantizers.149 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.down_proj_weight_quantizers.150 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.down_proj_weight_quantizers.151 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.down_proj_weight_quantizers.152 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.down_proj_weight_quantizers.153 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.down_proj_weight_quantizers.154 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.down_proj_weight_quantizers.155 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.58e-01 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.down_proj_weight_quantizers.156 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.down_proj_weight_quantizers.157 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.12e-01 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.down_proj_weight_quantizers.158 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.down_proj_weight_quantizers.159 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.down_proj_weight_quantizers.160 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.down_proj_weight_quantizers.161 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.down_proj_weight_quantizers.162 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.55e-01 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.down_proj_weight_quantizers.163 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.09e-01 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.down_proj_weight_quantizers.164 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.down_proj_weight_quantizers.165 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.down_proj_weight_quantizers.166 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.down_proj_weight_quantizers.167 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.45e-01 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.down_proj_weight_quantizers.168 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.down_proj_weight_quantizers.169 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.down_proj_weight_quantizers.170 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.18e-01 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.down_proj_weight_quantizers.171 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.93e-01 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.down_proj_weight_quantizers.172 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.down_proj_weight_quantizers.173 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.down_proj_weight_quantizers.174 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.down_proj_weight_quantizers.175 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.46e-01 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.down_proj_weight_quantizers.176 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.14e-01 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.down_proj_weight_quantizers.177 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.60e-01 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.down_proj_weight_quantizers.178 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.down_proj_weight_quantizers.179 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.down_proj_weight_quantizers.180 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.88e-01 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.down_proj_weight_quantizers.181 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.24e-01 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.down_proj_weight_quantizers.182 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.down_proj_weight_quantizers.183 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.17e-01 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.down_proj_weight_quantizers.184 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.64e-02 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.down_proj_weight_quantizers.185 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.20e-01 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.down_proj_weight_quantizers.186 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.59e-02 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.down_proj_weight_quantizers.187 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.down_proj_weight_quantizers.188 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.down_proj_weight_quantizers.189 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.down_proj_weight_quantizers.190 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.14e-01 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.down_proj_weight_quantizers.191 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.45e-02 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.down_proj_weight_quantizers.192 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.down_proj_weight_quantizers.193 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.down_proj_weight_quantizers.194 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.down_proj_weight_quantizers.195 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.down_proj_weight_quantizers.196 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.down_proj_weight_quantizers.197 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.down_proj_weight_quantizers.198 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.down_proj_weight_quantizers.199 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.down_proj_weight_quantizers.200 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.54e-02 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.down_proj_weight_quantizers.201 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.down_proj_weight_quantizers.202 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.down_proj_weight_quantizers.203 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.down_proj_weight_quantizers.204 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.49e-01 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.down_proj_weight_quantizers.205 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.down_proj_weight_quantizers.206 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.down_proj_weight_quantizers.207 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.20e-01 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.down_proj_weight_quantizers.208 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.down_proj_weight_quantizers.209 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.down_proj_weight_quantizers.210 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.down_proj_weight_quantizers.211 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.down_proj_weight_quantizers.212 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.down_proj_weight_quantizers.213 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.down_proj_weight_quantizers.214 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.down_proj_weight_quantizers.215 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.67e-01 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.down_proj_weight_quantizers.216 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.down_proj_weight_quantizers.217 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.down_proj_weight_quantizers.218 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.down_proj_weight_quantizers.219 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.06e-01 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.down_proj_weight_quantizers.220 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.down_proj_weight_quantizers.221 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.down_proj_weight_quantizers.222 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.down_proj_weight_quantizers.223 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.16e-01 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.down_proj_weight_quantizers.224 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.down_proj_weight_quantizers.225 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.down_proj_weight_quantizers.226 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.down_proj_weight_quantizers.227 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.down_proj_weight_quantizers.228 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.69e-01 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.down_proj_weight_quantizers.229 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.down_proj_weight_quantizers.230 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.down_proj_weight_quantizers.231 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.down_proj_weight_quantizers.232 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.down_proj_weight_quantizers.233 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.down_proj_weight_quantizers.234 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.down_proj_weight_quantizers.235 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.down_proj_weight_quantizers.236 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.down_proj_weight_quantizers.237 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.down_proj_weight_quantizers.238 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.down_proj_weight_quantizers.239 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.down_proj_weight_quantizers.240 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.down_proj_weight_quantizers.241 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.down_proj_weight_quantizers.242 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.down_proj_weight_quantizers.243 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.down_proj_weight_quantizers.244 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.down_proj_weight_quantizers.245 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.down_proj_weight_quantizers.246 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.down_proj_weight_quantizers.247 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.down_proj_weight_quantizers.248 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.34e-01 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.down_proj_weight_quantizers.249 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.59e-02 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.down_proj_weight_quantizers.250 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.down_proj_weight_quantizers.251 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.down_proj_weight_quantizers.252 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.52e-01 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.down_proj_weight_quantizers.253 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.down_proj_weight_quantizers.254 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.down_proj_weight_quantizers.255 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.64.mlp.shared_experts.gate_proj.input_quantizer TensorQuantizer(disabled) +model.layers.64.mlp.shared_experts.gate_proj.output_quantizer TensorQuantizer(disabled) +model.layers.64.mlp.shared_experts.gate_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.64.mlp.shared_experts.up_proj.input_quantizer TensorQuantizer(disabled) +model.layers.64.mlp.shared_experts.up_proj.output_quantizer TensorQuantizer(disabled) +model.layers.64.mlp.shared_experts.up_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.64.mlp.shared_experts.down_proj.input_quantizer TensorQuantizer(disabled) +model.layers.64.mlp.shared_experts.down_proj.output_quantizer TensorQuantizer(disabled) +model.layers.64.mlp.shared_experts.down_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.65.self_attn.q_a_proj.input_quantizer TensorQuantizer(disabled) +model.layers.65.self_attn.q_a_proj.output_quantizer TensorQuantizer(disabled) +model.layers.65.self_attn.q_a_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.65.self_attn.q_b_proj.input_quantizer TensorQuantizer(disabled) +model.layers.65.self_attn.q_b_proj.output_quantizer TensorQuantizer(disabled) +model.layers.65.self_attn.q_b_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.65.self_attn.kv_a_proj_with_mqa.input_quantizer TensorQuantizer(disabled) +model.layers.65.self_attn.kv_a_proj_with_mqa.output_quantizer TensorQuantizer(disabled) +model.layers.65.self_attn.kv_a_proj_with_mqa.weight_quantizer TensorQuantizer(disabled) +model.layers.65.self_attn.kv_b_proj.input_quantizer TensorQuantizer(disabled) +model.layers.65.self_attn.kv_b_proj.output_quantizer TensorQuantizer(disabled) +model.layers.65.self_attn.kv_b_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.65.self_attn.o_proj.input_quantizer TensorQuantizer(disabled) +model.layers.65.self_attn.o_proj.output_quantizer TensorQuantizer(disabled) +model.layers.65.self_attn.o_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.65.self_attn.q_bmm_quantizer TensorQuantizer(disabled) +model.layers.65.self_attn.k_bmm_quantizer TensorQuantizer((4, 3) bit fake per-tensor amax=4.48e+02(const) calibrator=MaxCalibrator quant) +model.layers.65.self_attn.v_bmm_quantizer TensorQuantizer((4, 3) bit fake per-tensor amax=4.48e+02(const) calibrator=MaxCalibrator quant) +model.layers.65.self_attn.p_bmm_quantizer TensorQuantizer(disabled) +model.layers.65.mlp.experts.gate_up_proj_input_quantizer TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.19e+00 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.gate_up_proj_weight_quantizers.0 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.gate_up_proj_weight_quantizers.1 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.gate_up_proj_weight_quantizers.2 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.61e-01 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.gate_up_proj_weight_quantizers.3 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.gate_up_proj_weight_quantizers.4 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.gate_up_proj_weight_quantizers.5 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.gate_up_proj_weight_quantizers.6 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.gate_up_proj_weight_quantizers.7 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.gate_up_proj_weight_quantizers.8 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.gate_up_proj_weight_quantizers.9 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.51e-01 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.gate_up_proj_weight_quantizers.10 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.gate_up_proj_weight_quantizers.11 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.gate_up_proj_weight_quantizers.12 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.gate_up_proj_weight_quantizers.13 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.gate_up_proj_weight_quantizers.14 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.gate_up_proj_weight_quantizers.15 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=6.25e-01 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.gate_up_proj_weight_quantizers.16 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.54e-01 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.gate_up_proj_weight_quantizers.17 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.gate_up_proj_weight_quantizers.18 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.gate_up_proj_weight_quantizers.19 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.55e-01 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.gate_up_proj_weight_quantizers.20 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.gate_up_proj_weight_quantizers.21 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.gate_up_proj_weight_quantizers.22 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.97e-01 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.gate_up_proj_weight_quantizers.23 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.45e-01 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.gate_up_proj_weight_quantizers.24 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.gate_up_proj_weight_quantizers.25 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.30e-01 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.gate_up_proj_weight_quantizers.26 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.gate_up_proj_weight_quantizers.27 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.13e-01 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.gate_up_proj_weight_quantizers.28 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.gate_up_proj_weight_quantizers.29 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.gate_up_proj_weight_quantizers.30 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.56e-01 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.gate_up_proj_weight_quantizers.31 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.gate_up_proj_weight_quantizers.32 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.58e-01 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.gate_up_proj_weight_quantizers.33 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.gate_up_proj_weight_quantizers.34 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.20e-01 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.gate_up_proj_weight_quantizers.35 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.gate_up_proj_weight_quantizers.36 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.gate_up_proj_weight_quantizers.37 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.gate_up_proj_weight_quantizers.38 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.gate_up_proj_weight_quantizers.39 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.gate_up_proj_weight_quantizers.40 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.84e-01 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.gate_up_proj_weight_quantizers.41 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.gate_up_proj_weight_quantizers.42 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.22e-01 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.gate_up_proj_weight_quantizers.43 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.gate_up_proj_weight_quantizers.44 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.gate_up_proj_weight_quantizers.45 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.94e-01 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.gate_up_proj_weight_quantizers.46 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.gate_up_proj_weight_quantizers.47 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.gate_up_proj_weight_quantizers.48 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.gate_up_proj_weight_quantizers.49 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.51e-01 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.gate_up_proj_weight_quantizers.50 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.gate_up_proj_weight_quantizers.51 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.14e-01 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.gate_up_proj_weight_quantizers.52 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.gate_up_proj_weight_quantizers.53 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.gate_up_proj_weight_quantizers.54 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.gate_up_proj_weight_quantizers.55 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.gate_up_proj_weight_quantizers.56 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.gate_up_proj_weight_quantizers.57 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.gate_up_proj_weight_quantizers.58 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.20e-01 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.gate_up_proj_weight_quantizers.59 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.61e-01 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.gate_up_proj_weight_quantizers.60 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.48e-01 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.gate_up_proj_weight_quantizers.61 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.gate_up_proj_weight_quantizers.62 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.gate_up_proj_weight_quantizers.63 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.37e-01 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.gate_up_proj_weight_quantizers.64 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=4.88e-01 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.gate_up_proj_weight_quantizers.65 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.gate_up_proj_weight_quantizers.66 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.gate_up_proj_weight_quantizers.67 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.gate_up_proj_weight_quantizers.68 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.gate_up_proj_weight_quantizers.69 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.03e-01 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.gate_up_proj_weight_quantizers.70 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.34e-01 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.gate_up_proj_weight_quantizers.71 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.gate_up_proj_weight_quantizers.72 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.gate_up_proj_weight_quantizers.73 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.gate_up_proj_weight_quantizers.74 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.16e-01 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.gate_up_proj_weight_quantizers.75 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.gate_up_proj_weight_quantizers.76 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.gate_up_proj_weight_quantizers.77 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.gate_up_proj_weight_quantizers.78 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.gate_up_proj_weight_quantizers.79 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.gate_up_proj_weight_quantizers.80 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.75e-01 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.gate_up_proj_weight_quantizers.81 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.gate_up_proj_weight_quantizers.82 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.gate_up_proj_weight_quantizers.83 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.gate_up_proj_weight_quantizers.84 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.19e-01 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.gate_up_proj_weight_quantizers.85 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.gate_up_proj_weight_quantizers.86 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.gate_up_proj_weight_quantizers.87 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.gate_up_proj_weight_quantizers.88 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.gate_up_proj_weight_quantizers.89 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.gate_up_proj_weight_quantizers.90 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.gate_up_proj_weight_quantizers.91 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.gate_up_proj_weight_quantizers.92 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.gate_up_proj_weight_quantizers.93 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.gate_up_proj_weight_quantizers.94 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.25e-01 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.gate_up_proj_weight_quantizers.95 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.24e-01 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.gate_up_proj_weight_quantizers.96 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.19e-01 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.gate_up_proj_weight_quantizers.97 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.gate_up_proj_weight_quantizers.98 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.gate_up_proj_weight_quantizers.99 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.36e-01 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.gate_up_proj_weight_quantizers.100 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.gate_up_proj_weight_quantizers.101 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.gate_up_proj_weight_quantizers.102 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.gate_up_proj_weight_quantizers.103 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.gate_up_proj_weight_quantizers.104 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.gate_up_proj_weight_quantizers.105 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.gate_up_proj_weight_quantizers.106 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.gate_up_proj_weight_quantizers.107 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.gate_up_proj_weight_quantizers.108 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.gate_up_proj_weight_quantizers.109 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.gate_up_proj_weight_quantizers.110 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.18e-01 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.gate_up_proj_weight_quantizers.111 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.gate_up_proj_weight_quantizers.112 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.gate_up_proj_weight_quantizers.113 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.gate_up_proj_weight_quantizers.114 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.gate_up_proj_weight_quantizers.115 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.04e-01 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.gate_up_proj_weight_quantizers.116 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.gate_up_proj_weight_quantizers.117 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.gate_up_proj_weight_quantizers.118 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.41e-01 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.gate_up_proj_weight_quantizers.119 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.gate_up_proj_weight_quantizers.120 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.24e-01 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.gate_up_proj_weight_quantizers.121 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.gate_up_proj_weight_quantizers.122 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.15e-01 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.gate_up_proj_weight_quantizers.123 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.gate_up_proj_weight_quantizers.124 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.gate_up_proj_weight_quantizers.125 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.32e-01 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.gate_up_proj_weight_quantizers.126 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.gate_up_proj_weight_quantizers.127 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.gate_up_proj_weight_quantizers.128 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.gate_up_proj_weight_quantizers.129 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.33e-01 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.gate_up_proj_weight_quantizers.130 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=4.88e-01 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.gate_up_proj_weight_quantizers.131 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.25e-01 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.gate_up_proj_weight_quantizers.132 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.19e-01 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.gate_up_proj_weight_quantizers.133 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.gate_up_proj_weight_quantizers.134 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.gate_up_proj_weight_quantizers.135 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.gate_up_proj_weight_quantizers.136 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.40e-01 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.gate_up_proj_weight_quantizers.137 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.gate_up_proj_weight_quantizers.138 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.gate_up_proj_weight_quantizers.139 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.21e-01 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.gate_up_proj_weight_quantizers.140 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.gate_up_proj_weight_quantizers.141 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.gate_up_proj_weight_quantizers.142 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.gate_up_proj_weight_quantizers.143 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.20e-01 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.gate_up_proj_weight_quantizers.144 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.gate_up_proj_weight_quantizers.145 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=4.34e-01 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.gate_up_proj_weight_quantizers.146 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.20e-01 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.gate_up_proj_weight_quantizers.147 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.gate_up_proj_weight_quantizers.148 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.gate_up_proj_weight_quantizers.149 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.gate_up_proj_weight_quantizers.150 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.gate_up_proj_weight_quantizers.151 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.gate_up_proj_weight_quantizers.152 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.gate_up_proj_weight_quantizers.153 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.gate_up_proj_weight_quantizers.154 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.gate_up_proj_weight_quantizers.155 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.gate_up_proj_weight_quantizers.156 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.gate_up_proj_weight_quantizers.157 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.gate_up_proj_weight_quantizers.158 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.22e-01 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.gate_up_proj_weight_quantizers.159 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.gate_up_proj_weight_quantizers.160 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.gate_up_proj_weight_quantizers.161 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.gate_up_proj_weight_quantizers.162 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.15e-01 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.gate_up_proj_weight_quantizers.163 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.gate_up_proj_weight_quantizers.164 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.gate_up_proj_weight_quantizers.165 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.gate_up_proj_weight_quantizers.166 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.gate_up_proj_weight_quantizers.167 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.gate_up_proj_weight_quantizers.168 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.gate_up_proj_weight_quantizers.169 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.gate_up_proj_weight_quantizers.170 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.gate_up_proj_weight_quantizers.171 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.17e-01 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.gate_up_proj_weight_quantizers.172 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.gate_up_proj_weight_quantizers.173 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.gate_up_proj_weight_quantizers.174 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.67e-01 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.gate_up_proj_weight_quantizers.175 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.gate_up_proj_weight_quantizers.176 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.13e-01 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.gate_up_proj_weight_quantizers.177 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.73e-01 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.gate_up_proj_weight_quantizers.178 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.gate_up_proj_weight_quantizers.179 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.gate_up_proj_weight_quantizers.180 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.gate_up_proj_weight_quantizers.181 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.gate_up_proj_weight_quantizers.182 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.gate_up_proj_weight_quantizers.183 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.gate_up_proj_weight_quantizers.184 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.gate_up_proj_weight_quantizers.185 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.gate_up_proj_weight_quantizers.186 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.gate_up_proj_weight_quantizers.187 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.gate_up_proj_weight_quantizers.188 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.gate_up_proj_weight_quantizers.189 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.gate_up_proj_weight_quantizers.190 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.gate_up_proj_weight_quantizers.191 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.gate_up_proj_weight_quantizers.192 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.20e-01 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.gate_up_proj_weight_quantizers.193 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.gate_up_proj_weight_quantizers.194 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.gate_up_proj_weight_quantizers.195 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.gate_up_proj_weight_quantizers.196 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.gate_up_proj_weight_quantizers.197 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.77e-01 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.gate_up_proj_weight_quantizers.198 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.gate_up_proj_weight_quantizers.199 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.gate_up_proj_weight_quantizers.200 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.gate_up_proj_weight_quantizers.201 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.gate_up_proj_weight_quantizers.202 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.gate_up_proj_weight_quantizers.203 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.gate_up_proj_weight_quantizers.204 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.gate_up_proj_weight_quantizers.205 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.32e-01 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.gate_up_proj_weight_quantizers.206 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.gate_up_proj_weight_quantizers.207 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.gate_up_proj_weight_quantizers.208 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.gate_up_proj_weight_quantizers.209 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.21e-01 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.gate_up_proj_weight_quantizers.210 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.47e-01 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.gate_up_proj_weight_quantizers.211 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.09e-01 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.gate_up_proj_weight_quantizers.212 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.gate_up_proj_weight_quantizers.213 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.gate_up_proj_weight_quantizers.214 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.gate_up_proj_weight_quantizers.215 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.gate_up_proj_weight_quantizers.216 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.gate_up_proj_weight_quantizers.217 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.gate_up_proj_weight_quantizers.218 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.gate_up_proj_weight_quantizers.219 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.gate_up_proj_weight_quantizers.220 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.gate_up_proj_weight_quantizers.221 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.gate_up_proj_weight_quantizers.222 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.56e-01 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.gate_up_proj_weight_quantizers.223 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.18e-01 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.gate_up_proj_weight_quantizers.224 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.14e-01 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.gate_up_proj_weight_quantizers.225 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.gate_up_proj_weight_quantizers.226 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.gate_up_proj_weight_quantizers.227 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.gate_up_proj_weight_quantizers.228 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.gate_up_proj_weight_quantizers.229 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.gate_up_proj_weight_quantizers.230 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.gate_up_proj_weight_quantizers.231 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.gate_up_proj_weight_quantizers.232 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.42e-01 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.gate_up_proj_weight_quantizers.233 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.gate_up_proj_weight_quantizers.234 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.21e-01 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.gate_up_proj_weight_quantizers.235 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.46e-01 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.gate_up_proj_weight_quantizers.236 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.gate_up_proj_weight_quantizers.237 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.gate_up_proj_weight_quantizers.238 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.68e-01 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.gate_up_proj_weight_quantizers.239 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=6.60e-01 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.gate_up_proj_weight_quantizers.240 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.gate_up_proj_weight_quantizers.241 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.gate_up_proj_weight_quantizers.242 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.gate_up_proj_weight_quantizers.243 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.gate_up_proj_weight_quantizers.244 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.26e-01 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.gate_up_proj_weight_quantizers.245 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.21e-01 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.gate_up_proj_weight_quantizers.246 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.18e-01 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.gate_up_proj_weight_quantizers.247 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.gate_up_proj_weight_quantizers.248 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.gate_up_proj_weight_quantizers.249 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.gate_up_proj_weight_quantizers.250 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.gate_up_proj_weight_quantizers.251 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.gate_up_proj_weight_quantizers.252 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.gate_up_proj_weight_quantizers.253 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.gate_up_proj_weight_quantizers.254 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.21e-01 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.gate_up_proj_weight_quantizers.255 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.58e-01 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.down_proj_input_quantizer TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=5.00e+01 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.down_proj_weight_quantizers.0 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.down_proj_weight_quantizers.1 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.50e-02 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.down_proj_weight_quantizers.2 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.48e-01 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.down_proj_weight_quantizers.3 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.down_proj_weight_quantizers.4 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.down_proj_weight_quantizers.5 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.59e-02 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.down_proj_weight_quantizers.6 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.down_proj_weight_quantizers.7 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.30e-01 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.down_proj_weight_quantizers.8 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.down_proj_weight_quantizers.9 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.down_proj_weight_quantizers.10 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.down_proj_weight_quantizers.11 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.down_proj_weight_quantizers.12 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.down_proj_weight_quantizers.13 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.down_proj_weight_quantizers.14 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.down_proj_weight_quantizers.15 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.24e-01 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.down_proj_weight_quantizers.16 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.52e-01 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.down_proj_weight_quantizers.17 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.down_proj_weight_quantizers.18 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.down_proj_weight_quantizers.19 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.down_proj_weight_quantizers.20 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.down_proj_weight_quantizers.21 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.down_proj_weight_quantizers.22 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.down_proj_weight_quantizers.23 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.down_proj_weight_quantizers.24 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.down_proj_weight_quantizers.25 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.down_proj_weight_quantizers.26 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.down_proj_weight_quantizers.27 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.down_proj_weight_quantizers.28 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.down_proj_weight_quantizers.29 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.54e-02 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.down_proj_weight_quantizers.30 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.down_proj_weight_quantizers.31 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.down_proj_weight_quantizers.32 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.down_proj_weight_quantizers.33 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.down_proj_weight_quantizers.34 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.30e-01 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.down_proj_weight_quantizers.35 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.down_proj_weight_quantizers.36 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.down_proj_weight_quantizers.37 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.down_proj_weight_quantizers.38 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.22e-01 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.down_proj_weight_quantizers.39 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.down_proj_weight_quantizers.40 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.down_proj_weight_quantizers.41 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.13e-01 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.down_proj_weight_quantizers.42 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.down_proj_weight_quantizers.43 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.down_proj_weight_quantizers.44 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.down_proj_weight_quantizers.45 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.down_proj_weight_quantizers.46 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.down_proj_weight_quantizers.47 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.down_proj_weight_quantizers.48 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.down_proj_weight_quantizers.49 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.down_proj_weight_quantizers.50 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.47e-01 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.down_proj_weight_quantizers.51 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.down_proj_weight_quantizers.52 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.down_proj_weight_quantizers.53 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.59e-02 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.down_proj_weight_quantizers.54 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.down_proj_weight_quantizers.55 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.down_proj_weight_quantizers.56 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.down_proj_weight_quantizers.57 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.down_proj_weight_quantizers.58 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.82e-01 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.down_proj_weight_quantizers.59 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.down_proj_weight_quantizers.60 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.down_proj_weight_quantizers.61 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.down_proj_weight_quantizers.62 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.down_proj_weight_quantizers.63 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.down_proj_weight_quantizers.64 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.71e-01 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.down_proj_weight_quantizers.65 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.down_proj_weight_quantizers.66 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.down_proj_weight_quantizers.67 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.down_proj_weight_quantizers.68 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.down_proj_weight_quantizers.69 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.41e-01 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.down_proj_weight_quantizers.70 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.down_proj_weight_quantizers.71 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.92e-01 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.down_proj_weight_quantizers.72 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.down_proj_weight_quantizers.73 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.down_proj_weight_quantizers.74 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.03e-01 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.down_proj_weight_quantizers.75 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.down_proj_weight_quantizers.76 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.down_proj_weight_quantizers.77 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.down_proj_weight_quantizers.78 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.59e-02 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.down_proj_weight_quantizers.79 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.down_proj_weight_quantizers.80 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.down_proj_weight_quantizers.81 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.down_proj_weight_quantizers.82 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.down_proj_weight_quantizers.83 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.down_proj_weight_quantizers.84 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.down_proj_weight_quantizers.85 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.down_proj_weight_quantizers.86 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.down_proj_weight_quantizers.87 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.down_proj_weight_quantizers.88 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.down_proj_weight_quantizers.89 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.down_proj_weight_quantizers.90 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.down_proj_weight_quantizers.91 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.down_proj_weight_quantizers.92 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.down_proj_weight_quantizers.93 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.down_proj_weight_quantizers.94 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.down_proj_weight_quantizers.95 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.down_proj_weight_quantizers.96 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.down_proj_weight_quantizers.97 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.down_proj_weight_quantizers.98 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.down_proj_weight_quantizers.99 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.down_proj_weight_quantizers.100 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.down_proj_weight_quantizers.101 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.down_proj_weight_quantizers.102 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.down_proj_weight_quantizers.103 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.down_proj_weight_quantizers.104 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.down_proj_weight_quantizers.105 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.down_proj_weight_quantizers.106 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.down_proj_weight_quantizers.107 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.down_proj_weight_quantizers.108 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.down_proj_weight_quantizers.109 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.down_proj_weight_quantizers.110 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.down_proj_weight_quantizers.111 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.down_proj_weight_quantizers.112 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.down_proj_weight_quantizers.113 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.down_proj_weight_quantizers.114 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.64e-02 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.down_proj_weight_quantizers.115 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.down_proj_weight_quantizers.116 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.down_proj_weight_quantizers.117 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.35e-02 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.down_proj_weight_quantizers.118 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.down_proj_weight_quantizers.119 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.down_proj_weight_quantizers.120 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.27e-01 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.down_proj_weight_quantizers.121 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.down_proj_weight_quantizers.122 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.down_proj_weight_quantizers.123 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.down_proj_weight_quantizers.124 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.down_proj_weight_quantizers.125 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.31e-01 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.down_proj_weight_quantizers.126 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.down_proj_weight_quantizers.127 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.down_proj_weight_quantizers.128 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.20e-01 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.down_proj_weight_quantizers.129 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.down_proj_weight_quantizers.130 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.down_proj_weight_quantizers.131 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.down_proj_weight_quantizers.132 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.down_proj_weight_quantizers.133 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.down_proj_weight_quantizers.134 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.down_proj_weight_quantizers.135 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.down_proj_weight_quantizers.136 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.10e-01 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.down_proj_weight_quantizers.137 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.down_proj_weight_quantizers.138 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.down_proj_weight_quantizers.139 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.down_proj_weight_quantizers.140 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.down_proj_weight_quantizers.141 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.down_proj_weight_quantizers.142 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.down_proj_weight_quantizers.143 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.down_proj_weight_quantizers.144 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.35e-01 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.down_proj_weight_quantizers.145 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.down_proj_weight_quantizers.146 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.17e-01 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.down_proj_weight_quantizers.147 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.down_proj_weight_quantizers.148 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.down_proj_weight_quantizers.149 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.down_proj_weight_quantizers.150 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.down_proj_weight_quantizers.151 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.down_proj_weight_quantizers.152 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.down_proj_weight_quantizers.153 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.down_proj_weight_quantizers.154 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.down_proj_weight_quantizers.155 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.down_proj_weight_quantizers.156 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.down_proj_weight_quantizers.157 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.down_proj_weight_quantizers.158 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.58e-01 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.down_proj_weight_quantizers.159 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.down_proj_weight_quantizers.160 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.down_proj_weight_quantizers.161 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.down_proj_weight_quantizers.162 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.down_proj_weight_quantizers.163 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.down_proj_weight_quantizers.164 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.down_proj_weight_quantizers.165 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.down_proj_weight_quantizers.166 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.down_proj_weight_quantizers.167 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.down_proj_weight_quantizers.168 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.down_proj_weight_quantizers.169 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.down_proj_weight_quantizers.170 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.11e-02 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.down_proj_weight_quantizers.171 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.down_proj_weight_quantizers.172 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.96e-01 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.down_proj_weight_quantizers.173 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.down_proj_weight_quantizers.174 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.down_proj_weight_quantizers.175 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.down_proj_weight_quantizers.176 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.48e-01 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.down_proj_weight_quantizers.177 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.down_proj_weight_quantizers.178 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.down_proj_weight_quantizers.179 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.down_proj_weight_quantizers.180 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.down_proj_weight_quantizers.181 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.down_proj_weight_quantizers.182 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.down_proj_weight_quantizers.183 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.down_proj_weight_quantizers.184 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.down_proj_weight_quantizers.185 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.down_proj_weight_quantizers.186 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.down_proj_weight_quantizers.187 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.down_proj_weight_quantizers.188 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.down_proj_weight_quantizers.189 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.down_proj_weight_quantizers.190 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.down_proj_weight_quantizers.191 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.down_proj_weight_quantizers.192 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.down_proj_weight_quantizers.193 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.down_proj_weight_quantizers.194 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.down_proj_weight_quantizers.195 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.down_proj_weight_quantizers.196 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.down_proj_weight_quantizers.197 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.down_proj_weight_quantizers.198 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.down_proj_weight_quantizers.199 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.down_proj_weight_quantizers.200 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.down_proj_weight_quantizers.201 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.down_proj_weight_quantizers.202 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.down_proj_weight_quantizers.203 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.down_proj_weight_quantizers.204 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.down_proj_weight_quantizers.205 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.54e-01 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.down_proj_weight_quantizers.206 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.down_proj_weight_quantizers.207 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.down_proj_weight_quantizers.208 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.down_proj_weight_quantizers.209 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.29e-01 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.down_proj_weight_quantizers.210 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.07e-01 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.down_proj_weight_quantizers.211 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.down_proj_weight_quantizers.212 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.down_proj_weight_quantizers.213 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.down_proj_weight_quantizers.214 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.down_proj_weight_quantizers.215 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.down_proj_weight_quantizers.216 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.down_proj_weight_quantizers.217 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.down_proj_weight_quantizers.218 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.down_proj_weight_quantizers.219 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.down_proj_weight_quantizers.220 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.down_proj_weight_quantizers.221 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.64e-02 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.down_proj_weight_quantizers.222 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.down_proj_weight_quantizers.223 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.down_proj_weight_quantizers.224 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.17e-01 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.down_proj_weight_quantizers.225 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.down_proj_weight_quantizers.226 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.down_proj_weight_quantizers.227 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.down_proj_weight_quantizers.228 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.24e-01 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.down_proj_weight_quantizers.229 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.down_proj_weight_quantizers.230 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.down_proj_weight_quantizers.231 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.down_proj_weight_quantizers.232 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.down_proj_weight_quantizers.233 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.down_proj_weight_quantizers.234 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.down_proj_weight_quantizers.235 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.down_proj_weight_quantizers.236 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.05e-01 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.down_proj_weight_quantizers.237 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.down_proj_weight_quantizers.238 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.57e-01 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.down_proj_weight_quantizers.239 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.98e-01 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.down_proj_weight_quantizers.240 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.down_proj_weight_quantizers.241 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.down_proj_weight_quantizers.242 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.down_proj_weight_quantizers.243 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.down_proj_weight_quantizers.244 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.down_proj_weight_quantizers.245 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.down_proj_weight_quantizers.246 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.58e-01 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.down_proj_weight_quantizers.247 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.down_proj_weight_quantizers.248 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.25e-01 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.down_proj_weight_quantizers.249 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.22e-01 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.down_proj_weight_quantizers.250 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.down_proj_weight_quantizers.251 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.down_proj_weight_quantizers.252 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.95e-01 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.down_proj_weight_quantizers.253 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.down_proj_weight_quantizers.254 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.down_proj_weight_quantizers.255 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.20e-01 calibrator=MaxCalibrator quant) +model.layers.65.mlp.shared_experts.gate_proj.input_quantizer TensorQuantizer(disabled) +model.layers.65.mlp.shared_experts.gate_proj.output_quantizer TensorQuantizer(disabled) +model.layers.65.mlp.shared_experts.gate_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.65.mlp.shared_experts.up_proj.input_quantizer TensorQuantizer(disabled) +model.layers.65.mlp.shared_experts.up_proj.output_quantizer TensorQuantizer(disabled) +model.layers.65.mlp.shared_experts.up_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.65.mlp.shared_experts.down_proj.input_quantizer TensorQuantizer(disabled) +model.layers.65.mlp.shared_experts.down_proj.output_quantizer TensorQuantizer(disabled) +model.layers.65.mlp.shared_experts.down_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.66.self_attn.q_a_proj.input_quantizer TensorQuantizer(disabled) +model.layers.66.self_attn.q_a_proj.output_quantizer TensorQuantizer(disabled) +model.layers.66.self_attn.q_a_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.66.self_attn.q_b_proj.input_quantizer TensorQuantizer(disabled) +model.layers.66.self_attn.q_b_proj.output_quantizer TensorQuantizer(disabled) +model.layers.66.self_attn.q_b_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.66.self_attn.kv_a_proj_with_mqa.input_quantizer TensorQuantizer(disabled) +model.layers.66.self_attn.kv_a_proj_with_mqa.output_quantizer TensorQuantizer(disabled) +model.layers.66.self_attn.kv_a_proj_with_mqa.weight_quantizer TensorQuantizer(disabled) +model.layers.66.self_attn.kv_b_proj.input_quantizer TensorQuantizer(disabled) +model.layers.66.self_attn.kv_b_proj.output_quantizer TensorQuantizer(disabled) +model.layers.66.self_attn.kv_b_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.66.self_attn.o_proj.input_quantizer TensorQuantizer(disabled) +model.layers.66.self_attn.o_proj.output_quantizer TensorQuantizer(disabled) +model.layers.66.self_attn.o_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.66.self_attn.indexer.wq_b.input_quantizer TensorQuantizer(disabled) +model.layers.66.self_attn.indexer.wq_b.output_quantizer TensorQuantizer(disabled) +model.layers.66.self_attn.indexer.wq_b.weight_quantizer TensorQuantizer(disabled) +model.layers.66.self_attn.indexer.wk.input_quantizer TensorQuantizer(disabled) +model.layers.66.self_attn.indexer.wk.output_quantizer TensorQuantizer(disabled) +model.layers.66.self_attn.indexer.wk.weight_quantizer TensorQuantizer(disabled) +model.layers.66.self_attn.indexer.k_norm.input_quantizer TensorQuantizer(disabled) +model.layers.66.self_attn.indexer.k_norm.output_quantizer TensorQuantizer(disabled) +model.layers.66.self_attn.indexer.weights_proj.input_quantizer TensorQuantizer(disabled) +model.layers.66.self_attn.indexer.weights_proj.output_quantizer TensorQuantizer(disabled) +model.layers.66.self_attn.indexer.weights_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.66.self_attn.q_bmm_quantizer TensorQuantizer(disabled) +model.layers.66.self_attn.k_bmm_quantizer TensorQuantizer((4, 3) bit fake per-tensor amax=4.48e+02(const) calibrator=MaxCalibrator quant) +model.layers.66.self_attn.v_bmm_quantizer TensorQuantizer((4, 3) bit fake per-tensor amax=4.48e+02(const) calibrator=MaxCalibrator quant) +model.layers.66.self_attn.p_bmm_quantizer TensorQuantizer(disabled) +model.layers.66.mlp.experts.gate_up_proj_input_quantizer TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.02e+00 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.gate_up_proj_weight_quantizers.0 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.gate_up_proj_weight_quantizers.1 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.46e-01 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.gate_up_proj_weight_quantizers.2 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.gate_up_proj_weight_quantizers.3 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.gate_up_proj_weight_quantizers.4 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.gate_up_proj_weight_quantizers.5 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.gate_up_proj_weight_quantizers.6 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.14e-01 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.gate_up_proj_weight_quantizers.7 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.gate_up_proj_weight_quantizers.8 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.gate_up_proj_weight_quantizers.9 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.gate_up_proj_weight_quantizers.10 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.gate_up_proj_weight_quantizers.11 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.gate_up_proj_weight_quantizers.12 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.gate_up_proj_weight_quantizers.13 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.gate_up_proj_weight_quantizers.14 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.88e-01 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.gate_up_proj_weight_quantizers.15 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.gate_up_proj_weight_quantizers.16 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.60e-01 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.gate_up_proj_weight_quantizers.17 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.62e-01 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.gate_up_proj_weight_quantizers.18 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.gate_up_proj_weight_quantizers.19 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.gate_up_proj_weight_quantizers.20 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.gate_up_proj_weight_quantizers.21 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.42e-01 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.gate_up_proj_weight_quantizers.22 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.gate_up_proj_weight_quantizers.23 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.21e-01 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.gate_up_proj_weight_quantizers.24 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.gate_up_proj_weight_quantizers.25 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=4.08e-01 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.gate_up_proj_weight_quantizers.26 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.gate_up_proj_weight_quantizers.27 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.gate_up_proj_weight_quantizers.28 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.gate_up_proj_weight_quantizers.29 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.27e-01 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.gate_up_proj_weight_quantizers.30 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.gate_up_proj_weight_quantizers.31 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.19e-01 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.gate_up_proj_weight_quantizers.32 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.18e-01 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.gate_up_proj_weight_quantizers.33 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.44e-01 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.gate_up_proj_weight_quantizers.34 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.gate_up_proj_weight_quantizers.35 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.gate_up_proj_weight_quantizers.36 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.09e-01 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.gate_up_proj_weight_quantizers.37 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.gate_up_proj_weight_quantizers.38 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.gate_up_proj_weight_quantizers.39 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.44e-01 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.gate_up_proj_weight_quantizers.40 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.53e-01 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.gate_up_proj_weight_quantizers.41 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.gate_up_proj_weight_quantizers.42 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.gate_up_proj_weight_quantizers.43 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.gate_up_proj_weight_quantizers.44 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.gate_up_proj_weight_quantizers.45 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.18e-01 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.gate_up_proj_weight_quantizers.46 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.gate_up_proj_weight_quantizers.47 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.25e-01 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.gate_up_proj_weight_quantizers.48 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.gate_up_proj_weight_quantizers.49 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.gate_up_proj_weight_quantizers.50 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.gate_up_proj_weight_quantizers.51 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.gate_up_proj_weight_quantizers.52 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.gate_up_proj_weight_quantizers.53 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.60e-01 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.gate_up_proj_weight_quantizers.54 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.52e-01 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.gate_up_proj_weight_quantizers.55 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.gate_up_proj_weight_quantizers.56 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.29e-01 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.gate_up_proj_weight_quantizers.57 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.gate_up_proj_weight_quantizers.58 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.gate_up_proj_weight_quantizers.59 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.gate_up_proj_weight_quantizers.60 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.gate_up_proj_weight_quantizers.61 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.gate_up_proj_weight_quantizers.62 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.gate_up_proj_weight_quantizers.63 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.51e-01 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.gate_up_proj_weight_quantizers.64 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.gate_up_proj_weight_quantizers.65 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.gate_up_proj_weight_quantizers.66 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.gate_up_proj_weight_quantizers.67 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.gate_up_proj_weight_quantizers.68 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.gate_up_proj_weight_quantizers.69 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.gate_up_proj_weight_quantizers.70 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.14e-01 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.gate_up_proj_weight_quantizers.71 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.20e-01 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.gate_up_proj_weight_quantizers.72 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.gate_up_proj_weight_quantizers.73 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.gate_up_proj_weight_quantizers.74 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.gate_up_proj_weight_quantizers.75 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.gate_up_proj_weight_quantizers.76 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.89e-01 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.gate_up_proj_weight_quantizers.77 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.31e-01 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.gate_up_proj_weight_quantizers.78 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.36e-01 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.gate_up_proj_weight_quantizers.79 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=4.36e-01 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.gate_up_proj_weight_quantizers.80 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.gate_up_proj_weight_quantizers.81 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.gate_up_proj_weight_quantizers.82 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.gate_up_proj_weight_quantizers.83 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.46e-01 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.gate_up_proj_weight_quantizers.84 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.16e-01 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.gate_up_proj_weight_quantizers.85 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.39e-01 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.gate_up_proj_weight_quantizers.86 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.gate_up_proj_weight_quantizers.87 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.27e-01 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.gate_up_proj_weight_quantizers.88 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.28e-01 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.gate_up_proj_weight_quantizers.89 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.gate_up_proj_weight_quantizers.90 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.gate_up_proj_weight_quantizers.91 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.gate_up_proj_weight_quantizers.92 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.41e-01 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.gate_up_proj_weight_quantizers.93 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=6.29e-01 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.gate_up_proj_weight_quantizers.94 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.gate_up_proj_weight_quantizers.95 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.20e-01 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.gate_up_proj_weight_quantizers.96 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.18e-01 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.gate_up_proj_weight_quantizers.97 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.17e-01 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.gate_up_proj_weight_quantizers.98 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.gate_up_proj_weight_quantizers.99 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.35e-01 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.gate_up_proj_weight_quantizers.100 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.gate_up_proj_weight_quantizers.101 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.gate_up_proj_weight_quantizers.102 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.gate_up_proj_weight_quantizers.103 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.gate_up_proj_weight_quantizers.104 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.gate_up_proj_weight_quantizers.105 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.39e-01 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.gate_up_proj_weight_quantizers.106 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.19e-01 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.gate_up_proj_weight_quantizers.107 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.gate_up_proj_weight_quantizers.108 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.gate_up_proj_weight_quantizers.109 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.91e-01 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.gate_up_proj_weight_quantizers.110 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.22e-01 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.gate_up_proj_weight_quantizers.111 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.gate_up_proj_weight_quantizers.112 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.gate_up_proj_weight_quantizers.113 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.gate_up_proj_weight_quantizers.114 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.gate_up_proj_weight_quantizers.115 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.gate_up_proj_weight_quantizers.116 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.gate_up_proj_weight_quantizers.117 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.gate_up_proj_weight_quantizers.118 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.gate_up_proj_weight_quantizers.119 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.gate_up_proj_weight_quantizers.120 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.gate_up_proj_weight_quantizers.121 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.gate_up_proj_weight_quantizers.122 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.gate_up_proj_weight_quantizers.123 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.gate_up_proj_weight_quantizers.124 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.92e-01 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.gate_up_proj_weight_quantizers.125 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.gate_up_proj_weight_quantizers.126 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.gate_up_proj_weight_quantizers.127 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=5.47e-01 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.gate_up_proj_weight_quantizers.128 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.23e-01 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.gate_up_proj_weight_quantizers.129 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.gate_up_proj_weight_quantizers.130 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.20e-01 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.gate_up_proj_weight_quantizers.131 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.gate_up_proj_weight_quantizers.132 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.gate_up_proj_weight_quantizers.133 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.gate_up_proj_weight_quantizers.134 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.14e-01 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.gate_up_proj_weight_quantizers.135 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.gate_up_proj_weight_quantizers.136 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.gate_up_proj_weight_quantizers.137 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.96e-01 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.gate_up_proj_weight_quantizers.138 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.gate_up_proj_weight_quantizers.139 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.52e-01 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.gate_up_proj_weight_quantizers.140 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.39e-01 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.gate_up_proj_weight_quantizers.141 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.23e-01 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.gate_up_proj_weight_quantizers.142 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.32e-01 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.gate_up_proj_weight_quantizers.143 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.gate_up_proj_weight_quantizers.144 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.15e-01 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.gate_up_proj_weight_quantizers.145 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.24e-01 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.gate_up_proj_weight_quantizers.146 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.gate_up_proj_weight_quantizers.147 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.31e-01 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.gate_up_proj_weight_quantizers.148 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.gate_up_proj_weight_quantizers.149 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.gate_up_proj_weight_quantizers.150 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.gate_up_proj_weight_quantizers.151 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.31e-01 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.gate_up_proj_weight_quantizers.152 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.24e-01 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.gate_up_proj_weight_quantizers.153 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.gate_up_proj_weight_quantizers.154 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.gate_up_proj_weight_quantizers.155 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.74e-01 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.gate_up_proj_weight_quantizers.156 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.25e-01 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.gate_up_proj_weight_quantizers.157 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.gate_up_proj_weight_quantizers.158 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.31e-01 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.gate_up_proj_weight_quantizers.159 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.gate_up_proj_weight_quantizers.160 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.gate_up_proj_weight_quantizers.161 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.57e-01 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.gate_up_proj_weight_quantizers.162 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.gate_up_proj_weight_quantizers.163 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.gate_up_proj_weight_quantizers.164 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.gate_up_proj_weight_quantizers.165 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.64e-01 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.gate_up_proj_weight_quantizers.166 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.gate_up_proj_weight_quantizers.167 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.gate_up_proj_weight_quantizers.168 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.gate_up_proj_weight_quantizers.169 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.gate_up_proj_weight_quantizers.170 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.gate_up_proj_weight_quantizers.171 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.gate_up_proj_weight_quantizers.172 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.26e-01 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.gate_up_proj_weight_quantizers.173 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.17e-01 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.gate_up_proj_weight_quantizers.174 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.gate_up_proj_weight_quantizers.175 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.24e-01 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.gate_up_proj_weight_quantizers.176 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.gate_up_proj_weight_quantizers.177 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.gate_up_proj_weight_quantizers.178 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.13e-01 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.gate_up_proj_weight_quantizers.179 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.gate_up_proj_weight_quantizers.180 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.17e-01 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.gate_up_proj_weight_quantizers.181 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.17e-01 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.gate_up_proj_weight_quantizers.182 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.gate_up_proj_weight_quantizers.183 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.10e-01 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.gate_up_proj_weight_quantizers.184 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=5.16e-01 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.gate_up_proj_weight_quantizers.185 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.gate_up_proj_weight_quantizers.186 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.gate_up_proj_weight_quantizers.187 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.gate_up_proj_weight_quantizers.188 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.89e-01 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.gate_up_proj_weight_quantizers.189 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.14e-01 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.gate_up_proj_weight_quantizers.190 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.13e-01 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.gate_up_proj_weight_quantizers.191 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.13e-01 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.gate_up_proj_weight_quantizers.192 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.gate_up_proj_weight_quantizers.193 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.gate_up_proj_weight_quantizers.194 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.gate_up_proj_weight_quantizers.195 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.gate_up_proj_weight_quantizers.196 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.20e-01 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.gate_up_proj_weight_quantizers.197 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.33e-01 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.gate_up_proj_weight_quantizers.198 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.gate_up_proj_weight_quantizers.199 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.26e-01 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.gate_up_proj_weight_quantizers.200 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.gate_up_proj_weight_quantizers.201 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.54e-01 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.gate_up_proj_weight_quantizers.202 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.18e-01 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.gate_up_proj_weight_quantizers.203 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.18e-01 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.gate_up_proj_weight_quantizers.204 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.gate_up_proj_weight_quantizers.205 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.gate_up_proj_weight_quantizers.206 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.27e-01 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.gate_up_proj_weight_quantizers.207 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.gate_up_proj_weight_quantizers.208 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.gate_up_proj_weight_quantizers.209 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.gate_up_proj_weight_quantizers.210 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.gate_up_proj_weight_quantizers.211 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.gate_up_proj_weight_quantizers.212 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.gate_up_proj_weight_quantizers.213 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.gate_up_proj_weight_quantizers.214 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.15e-01 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.gate_up_proj_weight_quantizers.215 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.gate_up_proj_weight_quantizers.216 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.gate_up_proj_weight_quantizers.217 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.gate_up_proj_weight_quantizers.218 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.62e-01 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.gate_up_proj_weight_quantizers.219 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.gate_up_proj_weight_quantizers.220 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.34e-01 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.gate_up_proj_weight_quantizers.221 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.88e-01 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.gate_up_proj_weight_quantizers.222 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.gate_up_proj_weight_quantizers.223 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.gate_up_proj_weight_quantizers.224 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.37e-01 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.gate_up_proj_weight_quantizers.225 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.gate_up_proj_weight_quantizers.226 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.21e-01 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.gate_up_proj_weight_quantizers.227 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.25e-01 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.gate_up_proj_weight_quantizers.228 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.gate_up_proj_weight_quantizers.229 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.gate_up_proj_weight_quantizers.230 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.24e-01 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.gate_up_proj_weight_quantizers.231 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.57e-01 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.gate_up_proj_weight_quantizers.232 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.07e-01 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.gate_up_proj_weight_quantizers.233 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.gate_up_proj_weight_quantizers.234 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.14e-01 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.gate_up_proj_weight_quantizers.235 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.gate_up_proj_weight_quantizers.236 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.gate_up_proj_weight_quantizers.237 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.gate_up_proj_weight_quantizers.238 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.gate_up_proj_weight_quantizers.239 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.gate_up_proj_weight_quantizers.240 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.gate_up_proj_weight_quantizers.241 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.gate_up_proj_weight_quantizers.242 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.gate_up_proj_weight_quantizers.243 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.gate_up_proj_weight_quantizers.244 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.gate_up_proj_weight_quantizers.245 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.gate_up_proj_weight_quantizers.246 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.gate_up_proj_weight_quantizers.247 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.34e-01 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.gate_up_proj_weight_quantizers.248 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.20e-01 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.gate_up_proj_weight_quantizers.249 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.gate_up_proj_weight_quantizers.250 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.gate_up_proj_weight_quantizers.251 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.48e-01 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.gate_up_proj_weight_quantizers.252 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.gate_up_proj_weight_quantizers.253 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.gate_up_proj_weight_quantizers.254 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.gate_up_proj_weight_quantizers.255 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.17e-01 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.down_proj_input_quantizer TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=4.70e+01 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.down_proj_weight_quantizers.0 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.20e-01 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.down_proj_weight_quantizers.1 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.down_proj_weight_quantizers.2 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.down_proj_weight_quantizers.3 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.down_proj_weight_quantizers.4 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.down_proj_weight_quantizers.5 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.down_proj_weight_quantizers.6 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.down_proj_weight_quantizers.7 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.down_proj_weight_quantizers.8 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.down_proj_weight_quantizers.9 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.down_proj_weight_quantizers.10 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.down_proj_weight_quantizers.11 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.down_proj_weight_quantizers.12 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.down_proj_weight_quantizers.13 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.down_proj_weight_quantizers.14 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.58e-01 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.down_proj_weight_quantizers.15 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.down_proj_weight_quantizers.16 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.30e-01 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.down_proj_weight_quantizers.17 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.16e-01 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.down_proj_weight_quantizers.18 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.down_proj_weight_quantizers.19 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.down_proj_weight_quantizers.20 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.down_proj_weight_quantizers.21 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.20e-01 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.down_proj_weight_quantizers.22 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.down_proj_weight_quantizers.23 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.down_proj_weight_quantizers.24 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.64e-02 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.down_proj_weight_quantizers.25 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.61e-01 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.down_proj_weight_quantizers.26 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.down_proj_weight_quantizers.27 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.down_proj_weight_quantizers.28 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.down_proj_weight_quantizers.29 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.down_proj_weight_quantizers.30 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.down_proj_weight_quantizers.31 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.down_proj_weight_quantizers.32 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.down_proj_weight_quantizers.33 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.down_proj_weight_quantizers.34 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.down_proj_weight_quantizers.35 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.down_proj_weight_quantizers.36 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.down_proj_weight_quantizers.37 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.down_proj_weight_quantizers.38 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.down_proj_weight_quantizers.39 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.down_proj_weight_quantizers.40 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.down_proj_weight_quantizers.41 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.down_proj_weight_quantizers.42 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.down_proj_weight_quantizers.43 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.64e-02 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.down_proj_weight_quantizers.44 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.82e-01 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.down_proj_weight_quantizers.45 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.down_proj_weight_quantizers.46 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.down_proj_weight_quantizers.47 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.34e-01 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.down_proj_weight_quantizers.48 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.down_proj_weight_quantizers.49 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.down_proj_weight_quantizers.50 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.down_proj_weight_quantizers.51 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.down_proj_weight_quantizers.52 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.down_proj_weight_quantizers.53 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.14e-01 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.down_proj_weight_quantizers.54 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.59e-02 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.down_proj_weight_quantizers.55 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.down_proj_weight_quantizers.56 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.down_proj_weight_quantizers.57 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.down_proj_weight_quantizers.58 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.down_proj_weight_quantizers.59 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.down_proj_weight_quantizers.60 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.down_proj_weight_quantizers.61 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.down_proj_weight_quantizers.62 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.down_proj_weight_quantizers.63 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.down_proj_weight_quantizers.64 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.down_proj_weight_quantizers.65 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.down_proj_weight_quantizers.66 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.down_proj_weight_quantizers.67 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.down_proj_weight_quantizers.68 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.down_proj_weight_quantizers.69 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.down_proj_weight_quantizers.70 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.down_proj_weight_quantizers.71 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.down_proj_weight_quantizers.72 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.down_proj_weight_quantizers.73 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.down_proj_weight_quantizers.74 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.down_proj_weight_quantizers.75 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.down_proj_weight_quantizers.76 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.down_proj_weight_quantizers.77 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.59e-02 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.down_proj_weight_quantizers.78 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.88e-01 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.down_proj_weight_quantizers.79 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.28e-01 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.down_proj_weight_quantizers.80 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.84e-01 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.down_proj_weight_quantizers.81 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.down_proj_weight_quantizers.82 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.down_proj_weight_quantizers.83 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.down_proj_weight_quantizers.84 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.26e-01 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.down_proj_weight_quantizers.85 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.58e-01 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.down_proj_weight_quantizers.86 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.down_proj_weight_quantizers.87 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.down_proj_weight_quantizers.88 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.16e-01 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.down_proj_weight_quantizers.89 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.down_proj_weight_quantizers.90 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.down_proj_weight_quantizers.91 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.down_proj_weight_quantizers.92 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.down_proj_weight_quantizers.93 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.34e-01 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.down_proj_weight_quantizers.94 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.down_proj_weight_quantizers.95 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.down_proj_weight_quantizers.96 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.down_proj_weight_quantizers.97 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.down_proj_weight_quantizers.98 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.down_proj_weight_quantizers.99 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.down_proj_weight_quantizers.100 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.down_proj_weight_quantizers.101 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.down_proj_weight_quantizers.102 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.down_proj_weight_quantizers.103 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.down_proj_weight_quantizers.104 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.down_proj_weight_quantizers.105 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.down_proj_weight_quantizers.106 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.down_proj_weight_quantizers.107 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.down_proj_weight_quantizers.108 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.down_proj_weight_quantizers.109 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.down_proj_weight_quantizers.110 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.down_proj_weight_quantizers.111 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.down_proj_weight_quantizers.112 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.32e-01 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.down_proj_weight_quantizers.113 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.down_proj_weight_quantizers.114 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.down_proj_weight_quantizers.115 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.down_proj_weight_quantizers.116 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.down_proj_weight_quantizers.117 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.down_proj_weight_quantizers.118 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.down_proj_weight_quantizers.119 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.down_proj_weight_quantizers.120 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.down_proj_weight_quantizers.121 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.59e-02 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.down_proj_weight_quantizers.122 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.down_proj_weight_quantizers.123 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.down_proj_weight_quantizers.124 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.45e-01 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.down_proj_weight_quantizers.125 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.down_proj_weight_quantizers.126 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.59e-02 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.down_proj_weight_quantizers.127 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.44e-01 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.down_proj_weight_quantizers.128 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.down_proj_weight_quantizers.129 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.down_proj_weight_quantizers.130 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.down_proj_weight_quantizers.131 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.down_proj_weight_quantizers.132 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.down_proj_weight_quantizers.133 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.down_proj_weight_quantizers.134 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.down_proj_weight_quantizers.135 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.down_proj_weight_quantizers.136 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.down_proj_weight_quantizers.137 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.down_proj_weight_quantizers.138 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.down_proj_weight_quantizers.139 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.down_proj_weight_quantizers.140 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.down_proj_weight_quantizers.141 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.down_proj_weight_quantizers.142 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.down_proj_weight_quantizers.143 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.down_proj_weight_quantizers.144 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.73e-01 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.down_proj_weight_quantizers.145 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.down_proj_weight_quantizers.146 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.down_proj_weight_quantizers.147 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.down_proj_weight_quantizers.148 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.down_proj_weight_quantizers.149 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.56e-01 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.down_proj_weight_quantizers.150 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.down_proj_weight_quantizers.151 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.31e-01 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.down_proj_weight_quantizers.152 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.35e-01 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.down_proj_weight_quantizers.153 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.down_proj_weight_quantizers.154 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.down_proj_weight_quantizers.155 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.50e-01 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.down_proj_weight_quantizers.156 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.down_proj_weight_quantizers.157 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.down_proj_weight_quantizers.158 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.down_proj_weight_quantizers.159 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.down_proj_weight_quantizers.160 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.down_proj_weight_quantizers.161 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.42e-01 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.down_proj_weight_quantizers.162 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.down_proj_weight_quantizers.163 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.down_proj_weight_quantizers.164 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.down_proj_weight_quantizers.165 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.down_proj_weight_quantizers.166 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.down_proj_weight_quantizers.167 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.down_proj_weight_quantizers.168 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.down_proj_weight_quantizers.169 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.down_proj_weight_quantizers.170 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.down_proj_weight_quantizers.171 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.down_proj_weight_quantizers.172 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.down_proj_weight_quantizers.173 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.27e-01 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.down_proj_weight_quantizers.174 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.down_proj_weight_quantizers.175 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.down_proj_weight_quantizers.176 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.down_proj_weight_quantizers.177 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.down_proj_weight_quantizers.178 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.down_proj_weight_quantizers.179 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.down_proj_weight_quantizers.180 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.down_proj_weight_quantizers.181 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.down_proj_weight_quantizers.182 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.down_proj_weight_quantizers.183 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.down_proj_weight_quantizers.184 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.18e-01 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.down_proj_weight_quantizers.185 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.40e-01 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.down_proj_weight_quantizers.186 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.down_proj_weight_quantizers.187 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.down_proj_weight_quantizers.188 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.down_proj_weight_quantizers.189 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.down_proj_weight_quantizers.190 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.down_proj_weight_quantizers.191 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.down_proj_weight_quantizers.192 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.down_proj_weight_quantizers.193 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.down_proj_weight_quantizers.194 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.down_proj_weight_quantizers.195 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.down_proj_weight_quantizers.196 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.down_proj_weight_quantizers.197 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.down_proj_weight_quantizers.198 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.down_proj_weight_quantizers.199 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.down_proj_weight_quantizers.200 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.down_proj_weight_quantizers.201 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.down_proj_weight_quantizers.202 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.down_proj_weight_quantizers.203 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.down_proj_weight_quantizers.204 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.down_proj_weight_quantizers.205 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.down_proj_weight_quantizers.206 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.down_proj_weight_quantizers.207 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.down_proj_weight_quantizers.208 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.down_proj_weight_quantizers.209 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.54e-02 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.down_proj_weight_quantizers.210 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.down_proj_weight_quantizers.211 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.down_proj_weight_quantizers.212 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.down_proj_weight_quantizers.213 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.down_proj_weight_quantizers.214 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.down_proj_weight_quantizers.215 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.down_proj_weight_quantizers.216 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.down_proj_weight_quantizers.217 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.down_proj_weight_quantizers.218 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.down_proj_weight_quantizers.219 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.down_proj_weight_quantizers.220 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.down_proj_weight_quantizers.221 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.59e-01 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.down_proj_weight_quantizers.222 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.down_proj_weight_quantizers.223 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.down_proj_weight_quantizers.224 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.down_proj_weight_quantizers.225 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.down_proj_weight_quantizers.226 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.50e-02 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.down_proj_weight_quantizers.227 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.down_proj_weight_quantizers.228 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.35e-02 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.down_proj_weight_quantizers.229 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.64e-02 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.down_proj_weight_quantizers.230 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.down_proj_weight_quantizers.231 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.down_proj_weight_quantizers.232 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.39e-01 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.down_proj_weight_quantizers.233 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.down_proj_weight_quantizers.234 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.down_proj_weight_quantizers.235 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.down_proj_weight_quantizers.236 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.down_proj_weight_quantizers.237 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.down_proj_weight_quantizers.238 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.down_proj_weight_quantizers.239 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.down_proj_weight_quantizers.240 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.down_proj_weight_quantizers.241 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.down_proj_weight_quantizers.242 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.down_proj_weight_quantizers.243 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.down_proj_weight_quantizers.244 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.down_proj_weight_quantizers.245 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.down_proj_weight_quantizers.246 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.down_proj_weight_quantizers.247 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.13e-01 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.down_proj_weight_quantizers.248 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.down_proj_weight_quantizers.249 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.down_proj_weight_quantizers.250 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.down_proj_weight_quantizers.251 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.30e-01 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.down_proj_weight_quantizers.252 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.down_proj_weight_quantizers.253 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.down_proj_weight_quantizers.254 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.down_proj_weight_quantizers.255 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.66.mlp.shared_experts.gate_proj.input_quantizer TensorQuantizer(disabled) +model.layers.66.mlp.shared_experts.gate_proj.output_quantizer TensorQuantizer(disabled) +model.layers.66.mlp.shared_experts.gate_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.66.mlp.shared_experts.up_proj.input_quantizer TensorQuantizer(disabled) +model.layers.66.mlp.shared_experts.up_proj.output_quantizer TensorQuantizer(disabled) +model.layers.66.mlp.shared_experts.up_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.66.mlp.shared_experts.down_proj.input_quantizer TensorQuantizer(disabled) +model.layers.66.mlp.shared_experts.down_proj.output_quantizer TensorQuantizer(disabled) +model.layers.66.mlp.shared_experts.down_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.67.self_attn.q_a_proj.input_quantizer TensorQuantizer(disabled) +model.layers.67.self_attn.q_a_proj.output_quantizer TensorQuantizer(disabled) +model.layers.67.self_attn.q_a_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.67.self_attn.q_b_proj.input_quantizer TensorQuantizer(disabled) +model.layers.67.self_attn.q_b_proj.output_quantizer TensorQuantizer(disabled) +model.layers.67.self_attn.q_b_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.67.self_attn.kv_a_proj_with_mqa.input_quantizer TensorQuantizer(disabled) +model.layers.67.self_attn.kv_a_proj_with_mqa.output_quantizer TensorQuantizer(disabled) +model.layers.67.self_attn.kv_a_proj_with_mqa.weight_quantizer TensorQuantizer(disabled) +model.layers.67.self_attn.kv_b_proj.input_quantizer TensorQuantizer(disabled) +model.layers.67.self_attn.kv_b_proj.output_quantizer TensorQuantizer(disabled) +model.layers.67.self_attn.kv_b_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.67.self_attn.o_proj.input_quantizer TensorQuantizer(disabled) +model.layers.67.self_attn.o_proj.output_quantizer TensorQuantizer(disabled) +model.layers.67.self_attn.o_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.67.self_attn.q_bmm_quantizer TensorQuantizer(disabled) +model.layers.67.self_attn.k_bmm_quantizer TensorQuantizer((4, 3) bit fake per-tensor amax=4.48e+02(const) calibrator=MaxCalibrator quant) +model.layers.67.self_attn.v_bmm_quantizer TensorQuantizer((4, 3) bit fake per-tensor amax=4.48e+02(const) calibrator=MaxCalibrator quant) +model.layers.67.self_attn.p_bmm_quantizer TensorQuantizer(disabled) +model.layers.67.mlp.experts.gate_up_proj_input_quantizer TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.98e+00 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.gate_up_proj_weight_quantizers.0 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.gate_up_proj_weight_quantizers.1 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.gate_up_proj_weight_quantizers.2 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.gate_up_proj_weight_quantizers.3 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.gate_up_proj_weight_quantizers.4 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.gate_up_proj_weight_quantizers.5 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.21e-01 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.gate_up_proj_weight_quantizers.6 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.gate_up_proj_weight_quantizers.7 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.gate_up_proj_weight_quantizers.8 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.18e-01 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.gate_up_proj_weight_quantizers.9 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.gate_up_proj_weight_quantizers.10 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.gate_up_proj_weight_quantizers.11 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.gate_up_proj_weight_quantizers.12 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.gate_up_proj_weight_quantizers.13 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.88e-01 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.gate_up_proj_weight_quantizers.14 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.gate_up_proj_weight_quantizers.15 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.19e-01 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.gate_up_proj_weight_quantizers.16 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.25e-01 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.gate_up_proj_weight_quantizers.17 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.gate_up_proj_weight_quantizers.18 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.gate_up_proj_weight_quantizers.19 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.58e-01 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.gate_up_proj_weight_quantizers.20 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.21e-01 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.gate_up_proj_weight_quantizers.21 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.89e-01 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.gate_up_proj_weight_quantizers.22 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=4.36e-01 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.gate_up_proj_weight_quantizers.23 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.gate_up_proj_weight_quantizers.24 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.gate_up_proj_weight_quantizers.25 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.49e-01 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.gate_up_proj_weight_quantizers.26 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.gate_up_proj_weight_quantizers.27 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.gate_up_proj_weight_quantizers.28 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.15e-01 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.gate_up_proj_weight_quantizers.29 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.gate_up_proj_weight_quantizers.30 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.gate_up_proj_weight_quantizers.31 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.gate_up_proj_weight_quantizers.32 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.gate_up_proj_weight_quantizers.33 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.gate_up_proj_weight_quantizers.34 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.87e-01 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.gate_up_proj_weight_quantizers.35 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.13e-01 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.gate_up_proj_weight_quantizers.36 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.gate_up_proj_weight_quantizers.37 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.gate_up_proj_weight_quantizers.38 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.14e-01 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.gate_up_proj_weight_quantizers.39 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.gate_up_proj_weight_quantizers.40 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.24e-01 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.gate_up_proj_weight_quantizers.41 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.gate_up_proj_weight_quantizers.42 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.gate_up_proj_weight_quantizers.43 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.gate_up_proj_weight_quantizers.44 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.gate_up_proj_weight_quantizers.45 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.gate_up_proj_weight_quantizers.46 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.gate_up_proj_weight_quantizers.47 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.59e-01 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.gate_up_proj_weight_quantizers.48 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.56e-01 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.gate_up_proj_weight_quantizers.49 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.gate_up_proj_weight_quantizers.50 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.gate_up_proj_weight_quantizers.51 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.gate_up_proj_weight_quantizers.52 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.21e-01 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.gate_up_proj_weight_quantizers.53 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.gate_up_proj_weight_quantizers.54 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.gate_up_proj_weight_quantizers.55 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.14e-01 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.gate_up_proj_weight_quantizers.56 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.gate_up_proj_weight_quantizers.57 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.gate_up_proj_weight_quantizers.58 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.gate_up_proj_weight_quantizers.59 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.71e-01 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.gate_up_proj_weight_quantizers.60 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.gate_up_proj_weight_quantizers.61 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.gate_up_proj_weight_quantizers.62 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.17e-01 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.gate_up_proj_weight_quantizers.63 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.gate_up_proj_weight_quantizers.64 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.20e-01 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.gate_up_proj_weight_quantizers.65 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.gate_up_proj_weight_quantizers.66 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.gate_up_proj_weight_quantizers.67 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.gate_up_proj_weight_quantizers.68 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.gate_up_proj_weight_quantizers.69 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.25e-01 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.gate_up_proj_weight_quantizers.70 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.gate_up_proj_weight_quantizers.71 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.gate_up_proj_weight_quantizers.72 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.gate_up_proj_weight_quantizers.73 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.gate_up_proj_weight_quantizers.74 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.gate_up_proj_weight_quantizers.75 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.gate_up_proj_weight_quantizers.76 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.gate_up_proj_weight_quantizers.77 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.85e-01 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.gate_up_proj_weight_quantizers.78 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.gate_up_proj_weight_quantizers.79 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.gate_up_proj_weight_quantizers.80 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.gate_up_proj_weight_quantizers.81 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.gate_up_proj_weight_quantizers.82 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.gate_up_proj_weight_quantizers.83 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.21e-01 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.gate_up_proj_weight_quantizers.84 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.gate_up_proj_weight_quantizers.85 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.gate_up_proj_weight_quantizers.86 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.gate_up_proj_weight_quantizers.87 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.gate_up_proj_weight_quantizers.88 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.14e-01 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.gate_up_proj_weight_quantizers.89 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.gate_up_proj_weight_quantizers.90 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.gate_up_proj_weight_quantizers.91 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.gate_up_proj_weight_quantizers.92 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.gate_up_proj_weight_quantizers.93 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.gate_up_proj_weight_quantizers.94 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.gate_up_proj_weight_quantizers.95 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.23e-01 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.gate_up_proj_weight_quantizers.96 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.gate_up_proj_weight_quantizers.97 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.25e-01 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.gate_up_proj_weight_quantizers.98 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.gate_up_proj_weight_quantizers.99 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.46e-01 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.gate_up_proj_weight_quantizers.100 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.66e-01 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.gate_up_proj_weight_quantizers.101 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.65e-01 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.gate_up_proj_weight_quantizers.102 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.28e-01 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.gate_up_proj_weight_quantizers.103 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.gate_up_proj_weight_quantizers.104 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.gate_up_proj_weight_quantizers.105 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.75e-01 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.gate_up_proj_weight_quantizers.106 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.gate_up_proj_weight_quantizers.107 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.14e-01 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.gate_up_proj_weight_quantizers.108 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=4.06e-01 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.gate_up_proj_weight_quantizers.109 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.18e-01 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.gate_up_proj_weight_quantizers.110 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.gate_up_proj_weight_quantizers.111 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.66e-01 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.gate_up_proj_weight_quantizers.112 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.15e-01 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.gate_up_proj_weight_quantizers.113 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.21e-01 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.gate_up_proj_weight_quantizers.114 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.gate_up_proj_weight_quantizers.115 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.43e-01 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.gate_up_proj_weight_quantizers.116 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.gate_up_proj_weight_quantizers.117 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.gate_up_proj_weight_quantizers.118 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.20e-01 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.gate_up_proj_weight_quantizers.119 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.gate_up_proj_weight_quantizers.120 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.gate_up_proj_weight_quantizers.121 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.13e-01 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.gate_up_proj_weight_quantizers.122 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.20e-01 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.gate_up_proj_weight_quantizers.123 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.gate_up_proj_weight_quantizers.124 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.gate_up_proj_weight_quantizers.125 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.17e-01 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.gate_up_proj_weight_quantizers.126 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.33e-01 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.gate_up_proj_weight_quantizers.127 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.76e-01 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.gate_up_proj_weight_quantizers.128 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.gate_up_proj_weight_quantizers.129 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.gate_up_proj_weight_quantizers.130 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.18e-01 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.gate_up_proj_weight_quantizers.131 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.22e-01 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.gate_up_proj_weight_quantizers.132 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.17e-01 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.gate_up_proj_weight_quantizers.133 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.16e-01 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.gate_up_proj_weight_quantizers.134 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.35e-01 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.gate_up_proj_weight_quantizers.135 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.gate_up_proj_weight_quantizers.136 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.18e-01 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.gate_up_proj_weight_quantizers.137 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.gate_up_proj_weight_quantizers.138 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.28e-01 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.gate_up_proj_weight_quantizers.139 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.gate_up_proj_weight_quantizers.140 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.gate_up_proj_weight_quantizers.141 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.90e-01 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.gate_up_proj_weight_quantizers.142 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.48e-01 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.gate_up_proj_weight_quantizers.143 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.gate_up_proj_weight_quantizers.144 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.gate_up_proj_weight_quantizers.145 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.gate_up_proj_weight_quantizers.146 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.21e-01 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.gate_up_proj_weight_quantizers.147 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.gate_up_proj_weight_quantizers.148 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.gate_up_proj_weight_quantizers.149 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.gate_up_proj_weight_quantizers.150 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.gate_up_proj_weight_quantizers.151 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.22e-01 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.gate_up_proj_weight_quantizers.152 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.gate_up_proj_weight_quantizers.153 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.gate_up_proj_weight_quantizers.154 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.gate_up_proj_weight_quantizers.155 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.gate_up_proj_weight_quantizers.156 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.98e-01 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.gate_up_proj_weight_quantizers.157 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.gate_up_proj_weight_quantizers.158 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.17e-01 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.gate_up_proj_weight_quantizers.159 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.gate_up_proj_weight_quantizers.160 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.23e-01 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.gate_up_proj_weight_quantizers.161 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.gate_up_proj_weight_quantizers.162 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.gate_up_proj_weight_quantizers.163 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.gate_up_proj_weight_quantizers.164 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.gate_up_proj_weight_quantizers.165 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.gate_up_proj_weight_quantizers.166 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.gate_up_proj_weight_quantizers.167 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.19e-01 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.gate_up_proj_weight_quantizers.168 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.gate_up_proj_weight_quantizers.169 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.gate_up_proj_weight_quantizers.170 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.gate_up_proj_weight_quantizers.171 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.gate_up_proj_weight_quantizers.172 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.24e-01 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.gate_up_proj_weight_quantizers.173 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.21e-01 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.gate_up_proj_weight_quantizers.174 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.gate_up_proj_weight_quantizers.175 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.99e-01 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.gate_up_proj_weight_quantizers.176 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.gate_up_proj_weight_quantizers.177 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.gate_up_proj_weight_quantizers.178 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.gate_up_proj_weight_quantizers.179 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.gate_up_proj_weight_quantizers.180 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.gate_up_proj_weight_quantizers.181 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.gate_up_proj_weight_quantizers.182 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.gate_up_proj_weight_quantizers.183 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.24e-01 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.gate_up_proj_weight_quantizers.184 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.16e-01 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.gate_up_proj_weight_quantizers.185 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.83e-01 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.gate_up_proj_weight_quantizers.186 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.42e-01 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.gate_up_proj_weight_quantizers.187 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.28e-01 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.gate_up_proj_weight_quantizers.188 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.gate_up_proj_weight_quantizers.189 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.09e-01 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.gate_up_proj_weight_quantizers.190 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.37e-01 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.gate_up_proj_weight_quantizers.191 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.13e-01 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.gate_up_proj_weight_quantizers.192 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.gate_up_proj_weight_quantizers.193 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.gate_up_proj_weight_quantizers.194 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.gate_up_proj_weight_quantizers.195 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.gate_up_proj_weight_quantizers.196 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=4.00e-01 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.gate_up_proj_weight_quantizers.197 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.gate_up_proj_weight_quantizers.198 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.gate_up_proj_weight_quantizers.199 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.17e-01 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.gate_up_proj_weight_quantizers.200 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.gate_up_proj_weight_quantizers.201 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.gate_up_proj_weight_quantizers.202 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.gate_up_proj_weight_quantizers.203 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.gate_up_proj_weight_quantizers.204 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.05e-01 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.gate_up_proj_weight_quantizers.205 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.gate_up_proj_weight_quantizers.206 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.gate_up_proj_weight_quantizers.207 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.22e-01 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.gate_up_proj_weight_quantizers.208 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.18e-01 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.gate_up_proj_weight_quantizers.209 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.gate_up_proj_weight_quantizers.210 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.26e-01 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.gate_up_proj_weight_quantizers.211 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.gate_up_proj_weight_quantizers.212 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.gate_up_proj_weight_quantizers.213 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.gate_up_proj_weight_quantizers.214 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.38e-01 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.gate_up_proj_weight_quantizers.215 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.gate_up_proj_weight_quantizers.216 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.gate_up_proj_weight_quantizers.217 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.82e-01 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.gate_up_proj_weight_quantizers.218 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.gate_up_proj_weight_quantizers.219 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.72e-01 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.gate_up_proj_weight_quantizers.220 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.gate_up_proj_weight_quantizers.221 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.24e-01 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.gate_up_proj_weight_quantizers.222 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.gate_up_proj_weight_quantizers.223 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=5.66e-01 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.gate_up_proj_weight_quantizers.224 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.gate_up_proj_weight_quantizers.225 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.gate_up_proj_weight_quantizers.226 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.68e-01 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.gate_up_proj_weight_quantizers.227 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.gate_up_proj_weight_quantizers.228 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.gate_up_proj_weight_quantizers.229 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.gate_up_proj_weight_quantizers.230 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.23e-01 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.gate_up_proj_weight_quantizers.231 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.30e-01 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.gate_up_proj_weight_quantizers.232 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.gate_up_proj_weight_quantizers.233 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.gate_up_proj_weight_quantizers.234 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.gate_up_proj_weight_quantizers.235 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.gate_up_proj_weight_quantizers.236 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.13e-01 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.gate_up_proj_weight_quantizers.237 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.36e-01 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.gate_up_proj_weight_quantizers.238 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.39e-01 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.gate_up_proj_weight_quantizers.239 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.gate_up_proj_weight_quantizers.240 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.gate_up_proj_weight_quantizers.241 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.gate_up_proj_weight_quantizers.242 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.65e-01 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.gate_up_proj_weight_quantizers.243 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.gate_up_proj_weight_quantizers.244 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.gate_up_proj_weight_quantizers.245 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.gate_up_proj_weight_quantizers.246 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=5.16e-01 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.gate_up_proj_weight_quantizers.247 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.gate_up_proj_weight_quantizers.248 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.gate_up_proj_weight_quantizers.249 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.gate_up_proj_weight_quantizers.250 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.gate_up_proj_weight_quantizers.251 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.16e-01 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.gate_up_proj_weight_quantizers.252 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.gate_up_proj_weight_quantizers.253 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.gate_up_proj_weight_quantizers.254 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.gate_up_proj_weight_quantizers.255 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.down_proj_input_quantizer TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=5.35e+01 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.down_proj_weight_quantizers.0 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.down_proj_weight_quantizers.1 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.down_proj_weight_quantizers.2 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.down_proj_weight_quantizers.3 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.down_proj_weight_quantizers.4 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.down_proj_weight_quantizers.5 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.48e-01 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.down_proj_weight_quantizers.6 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.37e-01 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.down_proj_weight_quantizers.7 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.down_proj_weight_quantizers.8 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.57e-01 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.down_proj_weight_quantizers.9 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.down_proj_weight_quantizers.10 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.down_proj_weight_quantizers.11 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.down_proj_weight_quantizers.12 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.down_proj_weight_quantizers.13 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.19e-01 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.down_proj_weight_quantizers.14 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.down_proj_weight_quantizers.15 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.52e-01 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.down_proj_weight_quantizers.16 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.73e-01 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.down_proj_weight_quantizers.17 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.down_proj_weight_quantizers.18 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.down_proj_weight_quantizers.19 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.down_proj_weight_quantizers.20 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.14e-01 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.down_proj_weight_quantizers.21 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.34e-01 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.down_proj_weight_quantizers.22 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.65e-01 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.down_proj_weight_quantizers.23 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.down_proj_weight_quantizers.24 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.down_proj_weight_quantizers.25 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.27e-01 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.down_proj_weight_quantizers.26 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.down_proj_weight_quantizers.27 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.down_proj_weight_quantizers.28 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.down_proj_weight_quantizers.29 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.down_proj_weight_quantizers.30 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.down_proj_weight_quantizers.31 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.down_proj_weight_quantizers.32 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.down_proj_weight_quantizers.33 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.down_proj_weight_quantizers.34 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.down_proj_weight_quantizers.35 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.down_proj_weight_quantizers.36 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.down_proj_weight_quantizers.37 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.down_proj_weight_quantizers.38 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.down_proj_weight_quantizers.39 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.down_proj_weight_quantizers.40 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.down_proj_weight_quantizers.41 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.down_proj_weight_quantizers.42 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.down_proj_weight_quantizers.43 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.down_proj_weight_quantizers.44 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.down_proj_weight_quantizers.45 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.down_proj_weight_quantizers.46 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.down_proj_weight_quantizers.47 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.down_proj_weight_quantizers.48 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.18e-01 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.down_proj_weight_quantizers.49 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.57e-01 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.down_proj_weight_quantizers.50 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.down_proj_weight_quantizers.51 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.down_proj_weight_quantizers.52 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.down_proj_weight_quantizers.53 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.down_proj_weight_quantizers.54 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.down_proj_weight_quantizers.55 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.down_proj_weight_quantizers.56 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.down_proj_weight_quantizers.57 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.down_proj_weight_quantizers.58 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.down_proj_weight_quantizers.59 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.64e-01 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.down_proj_weight_quantizers.60 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.down_proj_weight_quantizers.61 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.down_proj_weight_quantizers.62 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.down_proj_weight_quantizers.63 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.66e-01 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.down_proj_weight_quantizers.64 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.down_proj_weight_quantizers.65 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.down_proj_weight_quantizers.66 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.down_proj_weight_quantizers.67 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.down_proj_weight_quantizers.68 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.down_proj_weight_quantizers.69 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.18e-01 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.down_proj_weight_quantizers.70 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.down_proj_weight_quantizers.71 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.down_proj_weight_quantizers.72 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.33e-01 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.down_proj_weight_quantizers.73 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.down_proj_weight_quantizers.74 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.down_proj_weight_quantizers.75 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.down_proj_weight_quantizers.76 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.down_proj_weight_quantizers.77 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.down_proj_weight_quantizers.78 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.down_proj_weight_quantizers.79 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.down_proj_weight_quantizers.80 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.down_proj_weight_quantizers.81 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.down_proj_weight_quantizers.82 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.down_proj_weight_quantizers.83 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.down_proj_weight_quantizers.84 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.down_proj_weight_quantizers.85 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.down_proj_weight_quantizers.86 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.16e-01 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.down_proj_weight_quantizers.87 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.down_proj_weight_quantizers.88 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.down_proj_weight_quantizers.89 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.down_proj_weight_quantizers.90 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.down_proj_weight_quantizers.91 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.40e-02 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.down_proj_weight_quantizers.92 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.down_proj_weight_quantizers.93 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.down_proj_weight_quantizers.94 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.down_proj_weight_quantizers.95 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.down_proj_weight_quantizers.96 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.down_proj_weight_quantizers.97 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.down_proj_weight_quantizers.98 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.down_proj_weight_quantizers.99 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.down_proj_weight_quantizers.100 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.26e-01 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.down_proj_weight_quantizers.101 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.down_proj_weight_quantizers.102 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.82e-01 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.down_proj_weight_quantizers.103 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.down_proj_weight_quantizers.104 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.down_proj_weight_quantizers.105 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=4.39e-01 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.down_proj_weight_quantizers.106 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.down_proj_weight_quantizers.107 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.down_proj_weight_quantizers.108 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.19e-01 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.down_proj_weight_quantizers.109 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.22e-01 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.down_proj_weight_quantizers.110 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.21e-01 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.down_proj_weight_quantizers.111 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.34e-01 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.down_proj_weight_quantizers.112 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.down_proj_weight_quantizers.113 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.45e-01 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.down_proj_weight_quantizers.114 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.46e-01 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.down_proj_weight_quantizers.115 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.down_proj_weight_quantizers.116 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.down_proj_weight_quantizers.117 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.down_proj_weight_quantizers.118 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.down_proj_weight_quantizers.119 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.down_proj_weight_quantizers.120 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.down_proj_weight_quantizers.121 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.down_proj_weight_quantizers.122 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.down_proj_weight_quantizers.123 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.down_proj_weight_quantizers.124 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.down_proj_weight_quantizers.125 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.down_proj_weight_quantizers.126 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.down_proj_weight_quantizers.127 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.down_proj_weight_quantizers.128 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.down_proj_weight_quantizers.129 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.down_proj_weight_quantizers.130 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.down_proj_weight_quantizers.131 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.down_proj_weight_quantizers.132 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.down_proj_weight_quantizers.133 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.down_proj_weight_quantizers.134 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.13e-01 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.down_proj_weight_quantizers.135 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.down_proj_weight_quantizers.136 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.down_proj_weight_quantizers.137 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.down_proj_weight_quantizers.138 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.down_proj_weight_quantizers.139 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.58e-01 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.down_proj_weight_quantizers.140 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.down_proj_weight_quantizers.141 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.27e-01 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.down_proj_weight_quantizers.142 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.15e-01 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.down_proj_weight_quantizers.143 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.down_proj_weight_quantizers.144 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.down_proj_weight_quantizers.145 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.down_proj_weight_quantizers.146 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.down_proj_weight_quantizers.147 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.down_proj_weight_quantizers.148 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.down_proj_weight_quantizers.149 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.59e-02 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.down_proj_weight_quantizers.150 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.down_proj_weight_quantizers.151 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.down_proj_weight_quantizers.152 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.down_proj_weight_quantizers.153 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.down_proj_weight_quantizers.154 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.64e-02 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.down_proj_weight_quantizers.155 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.down_proj_weight_quantizers.156 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=5.08e-01 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.down_proj_weight_quantizers.157 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.down_proj_weight_quantizers.158 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.down_proj_weight_quantizers.159 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.33e-01 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.down_proj_weight_quantizers.160 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.down_proj_weight_quantizers.161 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.down_proj_weight_quantizers.162 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.14e-01 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.down_proj_weight_quantizers.163 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.down_proj_weight_quantizers.164 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.down_proj_weight_quantizers.165 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.down_proj_weight_quantizers.166 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.down_proj_weight_quantizers.167 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.down_proj_weight_quantizers.168 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.down_proj_weight_quantizers.169 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.down_proj_weight_quantizers.170 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.down_proj_weight_quantizers.171 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.down_proj_weight_quantizers.172 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.down_proj_weight_quantizers.173 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.27e-01 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.down_proj_weight_quantizers.174 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.down_proj_weight_quantizers.175 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.down_proj_weight_quantizers.176 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.down_proj_weight_quantizers.177 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.down_proj_weight_quantizers.178 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.down_proj_weight_quantizers.179 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.18e-01 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.down_proj_weight_quantizers.180 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.down_proj_weight_quantizers.181 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.down_proj_weight_quantizers.182 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.17e-01 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.down_proj_weight_quantizers.183 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.down_proj_weight_quantizers.184 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.down_proj_weight_quantizers.185 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.down_proj_weight_quantizers.186 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.down_proj_weight_quantizers.187 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.down_proj_weight_quantizers.188 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.47e-01 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.down_proj_weight_quantizers.189 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.down_proj_weight_quantizers.190 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.37e-01 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.down_proj_weight_quantizers.191 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.down_proj_weight_quantizers.192 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.down_proj_weight_quantizers.193 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.down_proj_weight_quantizers.194 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.down_proj_weight_quantizers.195 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.down_proj_weight_quantizers.196 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.25e-01 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.down_proj_weight_quantizers.197 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.down_proj_weight_quantizers.198 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.down_proj_weight_quantizers.199 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.down_proj_weight_quantizers.200 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.down_proj_weight_quantizers.201 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.down_proj_weight_quantizers.202 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.18e-01 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.down_proj_weight_quantizers.203 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.down_proj_weight_quantizers.204 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.16e-01 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.down_proj_weight_quantizers.205 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.down_proj_weight_quantizers.206 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.down_proj_weight_quantizers.207 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.down_proj_weight_quantizers.208 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.down_proj_weight_quantizers.209 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.down_proj_weight_quantizers.210 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.down_proj_weight_quantizers.211 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.down_proj_weight_quantizers.212 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.down_proj_weight_quantizers.213 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.down_proj_weight_quantizers.214 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.68e-01 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.down_proj_weight_quantizers.215 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.down_proj_weight_quantizers.216 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.down_proj_weight_quantizers.217 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.13e-01 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.down_proj_weight_quantizers.218 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.down_proj_weight_quantizers.219 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.down_proj_weight_quantizers.220 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.down_proj_weight_quantizers.221 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.down_proj_weight_quantizers.222 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.down_proj_weight_quantizers.223 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.46e-01 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.down_proj_weight_quantizers.224 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.down_proj_weight_quantizers.225 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.down_proj_weight_quantizers.226 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.down_proj_weight_quantizers.227 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.down_proj_weight_quantizers.228 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.down_proj_weight_quantizers.229 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.down_proj_weight_quantizers.230 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.down_proj_weight_quantizers.231 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.down_proj_weight_quantizers.232 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.54e-02 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.down_proj_weight_quantizers.233 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.77e-01 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.down_proj_weight_quantizers.234 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.down_proj_weight_quantizers.235 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.down_proj_weight_quantizers.236 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.down_proj_weight_quantizers.237 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.down_proj_weight_quantizers.238 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.down_proj_weight_quantizers.239 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.down_proj_weight_quantizers.240 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.down_proj_weight_quantizers.241 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.down_proj_weight_quantizers.242 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.17e-01 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.down_proj_weight_quantizers.243 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.down_proj_weight_quantizers.244 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.down_proj_weight_quantizers.245 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.down_proj_weight_quantizers.246 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.40e-01 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.down_proj_weight_quantizers.247 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.down_proj_weight_quantizers.248 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.down_proj_weight_quantizers.249 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.down_proj_weight_quantizers.250 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.down_proj_weight_quantizers.251 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.down_proj_weight_quantizers.252 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.down_proj_weight_quantizers.253 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.down_proj_weight_quantizers.254 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.down_proj_weight_quantizers.255 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.67.mlp.shared_experts.gate_proj.input_quantizer TensorQuantizer(disabled) +model.layers.67.mlp.shared_experts.gate_proj.output_quantizer TensorQuantizer(disabled) +model.layers.67.mlp.shared_experts.gate_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.67.mlp.shared_experts.up_proj.input_quantizer TensorQuantizer(disabled) +model.layers.67.mlp.shared_experts.up_proj.output_quantizer TensorQuantizer(disabled) +model.layers.67.mlp.shared_experts.up_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.67.mlp.shared_experts.down_proj.input_quantizer TensorQuantizer(disabled) +model.layers.67.mlp.shared_experts.down_proj.output_quantizer TensorQuantizer(disabled) +model.layers.67.mlp.shared_experts.down_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.68.self_attn.q_a_proj.input_quantizer TensorQuantizer(disabled) +model.layers.68.self_attn.q_a_proj.output_quantizer TensorQuantizer(disabled) +model.layers.68.self_attn.q_a_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.68.self_attn.q_b_proj.input_quantizer TensorQuantizer(disabled) +model.layers.68.self_attn.q_b_proj.output_quantizer TensorQuantizer(disabled) +model.layers.68.self_attn.q_b_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.68.self_attn.kv_a_proj_with_mqa.input_quantizer TensorQuantizer(disabled) +model.layers.68.self_attn.kv_a_proj_with_mqa.output_quantizer TensorQuantizer(disabled) +model.layers.68.self_attn.kv_a_proj_with_mqa.weight_quantizer TensorQuantizer(disabled) +model.layers.68.self_attn.kv_b_proj.input_quantizer TensorQuantizer(disabled) +model.layers.68.self_attn.kv_b_proj.output_quantizer TensorQuantizer(disabled) +model.layers.68.self_attn.kv_b_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.68.self_attn.o_proj.input_quantizer TensorQuantizer(disabled) +model.layers.68.self_attn.o_proj.output_quantizer TensorQuantizer(disabled) +model.layers.68.self_attn.o_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.68.self_attn.q_bmm_quantizer TensorQuantizer(disabled) +model.layers.68.self_attn.k_bmm_quantizer TensorQuantizer((4, 3) bit fake per-tensor amax=4.48e+02(const) calibrator=MaxCalibrator quant) +model.layers.68.self_attn.v_bmm_quantizer TensorQuantizer((4, 3) bit fake per-tensor amax=4.48e+02(const) calibrator=MaxCalibrator quant) +model.layers.68.self_attn.p_bmm_quantizer TensorQuantizer(disabled) +model.layers.68.mlp.experts.gate_up_proj_input_quantizer TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.14e+00 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.gate_up_proj_weight_quantizers.0 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.68e-01 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.gate_up_proj_weight_quantizers.1 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.gate_up_proj_weight_quantizers.2 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.23e-01 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.gate_up_proj_weight_quantizers.3 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.gate_up_proj_weight_quantizers.4 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.13e-01 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.gate_up_proj_weight_quantizers.5 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.16e-01 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.gate_up_proj_weight_quantizers.6 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.gate_up_proj_weight_quantizers.7 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.gate_up_proj_weight_quantizers.8 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.21e-01 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.gate_up_proj_weight_quantizers.9 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.89e-01 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.gate_up_proj_weight_quantizers.10 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.gate_up_proj_weight_quantizers.11 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.14e-01 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.gate_up_proj_weight_quantizers.12 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.20e-01 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.gate_up_proj_weight_quantizers.13 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.gate_up_proj_weight_quantizers.14 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.gate_up_proj_weight_quantizers.15 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.gate_up_proj_weight_quantizers.16 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.gate_up_proj_weight_quantizers.17 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.35e-01 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.gate_up_proj_weight_quantizers.18 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.gate_up_proj_weight_quantizers.19 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.gate_up_proj_weight_quantizers.20 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.69e-01 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.gate_up_proj_weight_quantizers.21 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.gate_up_proj_weight_quantizers.22 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.gate_up_proj_weight_quantizers.23 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.18e-01 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.gate_up_proj_weight_quantizers.24 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.27e-01 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.gate_up_proj_weight_quantizers.25 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.27e-01 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.gate_up_proj_weight_quantizers.26 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.gate_up_proj_weight_quantizers.27 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.18e+00 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.gate_up_proj_weight_quantizers.28 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.gate_up_proj_weight_quantizers.29 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.35e-01 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.gate_up_proj_weight_quantizers.30 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.gate_up_proj_weight_quantizers.31 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.gate_up_proj_weight_quantizers.32 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.89e-01 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.gate_up_proj_weight_quantizers.33 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.gate_up_proj_weight_quantizers.34 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.44e-01 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.gate_up_proj_weight_quantizers.35 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.15e-01 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.gate_up_proj_weight_quantizers.36 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.91e-01 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.gate_up_proj_weight_quantizers.37 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.gate_up_proj_weight_quantizers.38 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.gate_up_proj_weight_quantizers.39 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.16e-01 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.gate_up_proj_weight_quantizers.40 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.34e-01 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.gate_up_proj_weight_quantizers.41 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.gate_up_proj_weight_quantizers.42 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.gate_up_proj_weight_quantizers.43 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.98e-01 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.gate_up_proj_weight_quantizers.44 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.38e-01 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.gate_up_proj_weight_quantizers.45 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.96e-01 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.gate_up_proj_weight_quantizers.46 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.gate_up_proj_weight_quantizers.47 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.gate_up_proj_weight_quantizers.48 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.75e-01 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.gate_up_proj_weight_quantizers.49 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.gate_up_proj_weight_quantizers.50 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.46e-01 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.gate_up_proj_weight_quantizers.51 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.gate_up_proj_weight_quantizers.52 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.gate_up_proj_weight_quantizers.53 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.29e-01 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.gate_up_proj_weight_quantizers.54 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.73e+00 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.gate_up_proj_weight_quantizers.55 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.gate_up_proj_weight_quantizers.56 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.14e-01 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.gate_up_proj_weight_quantizers.57 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.gate_up_proj_weight_quantizers.58 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=5.82e-01 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.gate_up_proj_weight_quantizers.59 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.gate_up_proj_weight_quantizers.60 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.53e-01 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.gate_up_proj_weight_quantizers.61 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.15e-01 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.gate_up_proj_weight_quantizers.62 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.gate_up_proj_weight_quantizers.63 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.gate_up_proj_weight_quantizers.64 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.gate_up_proj_weight_quantizers.65 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.gate_up_proj_weight_quantizers.66 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.gate_up_proj_weight_quantizers.67 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.gate_up_proj_weight_quantizers.68 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.40e-01 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.gate_up_proj_weight_quantizers.69 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.gate_up_proj_weight_quantizers.70 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.gate_up_proj_weight_quantizers.71 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.75e-01 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.gate_up_proj_weight_quantizers.72 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.gate_up_proj_weight_quantizers.73 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.24e-01 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.gate_up_proj_weight_quantizers.74 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.gate_up_proj_weight_quantizers.75 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.15e-01 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.gate_up_proj_weight_quantizers.76 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.gate_up_proj_weight_quantizers.77 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.gate_up_proj_weight_quantizers.78 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.gate_up_proj_weight_quantizers.79 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.gate_up_proj_weight_quantizers.80 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.21e-01 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.gate_up_proj_weight_quantizers.81 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.gate_up_proj_weight_quantizers.82 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.gate_up_proj_weight_quantizers.83 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.gate_up_proj_weight_quantizers.84 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.86e-01 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.gate_up_proj_weight_quantizers.85 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.gate_up_proj_weight_quantizers.86 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.gate_up_proj_weight_quantizers.87 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=6.02e-01 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.gate_up_proj_weight_quantizers.88 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.gate_up_proj_weight_quantizers.89 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.64e-01 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.gate_up_proj_weight_quantizers.90 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.21e-01 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.gate_up_proj_weight_quantizers.91 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.gate_up_proj_weight_quantizers.92 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.gate_up_proj_weight_quantizers.93 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.gate_up_proj_weight_quantizers.94 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.gate_up_proj_weight_quantizers.95 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.gate_up_proj_weight_quantizers.96 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.48e-01 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.gate_up_proj_weight_quantizers.97 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.gate_up_proj_weight_quantizers.98 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.gate_up_proj_weight_quantizers.99 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.gate_up_proj_weight_quantizers.100 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.gate_up_proj_weight_quantizers.101 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.19e-01 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.gate_up_proj_weight_quantizers.102 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.74e-01 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.gate_up_proj_weight_quantizers.103 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.gate_up_proj_weight_quantizers.104 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.33e-01 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.gate_up_proj_weight_quantizers.105 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.gate_up_proj_weight_quantizers.106 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.gate_up_proj_weight_quantizers.107 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.gate_up_proj_weight_quantizers.108 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.gate_up_proj_weight_quantizers.109 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.gate_up_proj_weight_quantizers.110 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.15e-01 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.gate_up_proj_weight_quantizers.111 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.29e-01 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.gate_up_proj_weight_quantizers.112 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.gate_up_proj_weight_quantizers.113 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.54e-01 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.gate_up_proj_weight_quantizers.114 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.gate_up_proj_weight_quantizers.115 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.29e-01 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.gate_up_proj_weight_quantizers.116 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.gate_up_proj_weight_quantizers.117 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.gate_up_proj_weight_quantizers.118 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.18e-01 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.gate_up_proj_weight_quantizers.119 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.43e-01 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.gate_up_proj_weight_quantizers.120 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.gate_up_proj_weight_quantizers.121 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.gate_up_proj_weight_quantizers.122 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.gate_up_proj_weight_quantizers.123 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.gate_up_proj_weight_quantizers.124 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.17e-01 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.gate_up_proj_weight_quantizers.125 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.19e-01 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.gate_up_proj_weight_quantizers.126 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.23e-01 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.gate_up_proj_weight_quantizers.127 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.30e-01 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.gate_up_proj_weight_quantizers.128 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.16e-01 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.gate_up_proj_weight_quantizers.129 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.gate_up_proj_weight_quantizers.130 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.gate_up_proj_weight_quantizers.131 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.gate_up_proj_weight_quantizers.132 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.gate_up_proj_weight_quantizers.133 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.gate_up_proj_weight_quantizers.134 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.gate_up_proj_weight_quantizers.135 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.gate_up_proj_weight_quantizers.136 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.gate_up_proj_weight_quantizers.137 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.gate_up_proj_weight_quantizers.138 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.gate_up_proj_weight_quantizers.139 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.59e-01 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.gate_up_proj_weight_quantizers.140 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.gate_up_proj_weight_quantizers.141 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.gate_up_proj_weight_quantizers.142 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=4.57e-01 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.gate_up_proj_weight_quantizers.143 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.gate_up_proj_weight_quantizers.144 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.gate_up_proj_weight_quantizers.145 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.gate_up_proj_weight_quantizers.146 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.16e-01 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.gate_up_proj_weight_quantizers.147 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.gate_up_proj_weight_quantizers.148 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.gate_up_proj_weight_quantizers.149 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.gate_up_proj_weight_quantizers.150 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.22e-01 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.gate_up_proj_weight_quantizers.151 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.gate_up_proj_weight_quantizers.152 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.23e-01 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.gate_up_proj_weight_quantizers.153 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.15e-01 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.gate_up_proj_weight_quantizers.154 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.gate_up_proj_weight_quantizers.155 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.gate_up_proj_weight_quantizers.156 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.gate_up_proj_weight_quantizers.157 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.14e-01 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.gate_up_proj_weight_quantizers.158 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.gate_up_proj_weight_quantizers.159 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.gate_up_proj_weight_quantizers.160 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.44e-01 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.gate_up_proj_weight_quantizers.161 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.gate_up_proj_weight_quantizers.162 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.13e-01 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.gate_up_proj_weight_quantizers.163 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.37e-01 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.gate_up_proj_weight_quantizers.164 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.gate_up_proj_weight_quantizers.165 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.gate_up_proj_weight_quantizers.166 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.gate_up_proj_weight_quantizers.167 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=4.51e-01 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.gate_up_proj_weight_quantizers.168 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.gate_up_proj_weight_quantizers.169 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.54e-01 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.gate_up_proj_weight_quantizers.170 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.gate_up_proj_weight_quantizers.171 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.80e-01 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.gate_up_proj_weight_quantizers.172 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.gate_up_proj_weight_quantizers.173 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.20e-01 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.gate_up_proj_weight_quantizers.174 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.gate_up_proj_weight_quantizers.175 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.16e-01 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.gate_up_proj_weight_quantizers.176 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.23e-01 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.gate_up_proj_weight_quantizers.177 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.gate_up_proj_weight_quantizers.178 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.69e-01 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.gate_up_proj_weight_quantizers.179 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.74e-01 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.gate_up_proj_weight_quantizers.180 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.28e-01 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.gate_up_proj_weight_quantizers.181 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.34e-01 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.gate_up_proj_weight_quantizers.182 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.69e-01 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.gate_up_proj_weight_quantizers.183 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.25e-01 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.gate_up_proj_weight_quantizers.184 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.gate_up_proj_weight_quantizers.185 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.gate_up_proj_weight_quantizers.186 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.gate_up_proj_weight_quantizers.187 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.30e-01 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.gate_up_proj_weight_quantizers.188 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.19e-01 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.gate_up_proj_weight_quantizers.189 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.21e-01 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.gate_up_proj_weight_quantizers.190 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.gate_up_proj_weight_quantizers.191 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.gate_up_proj_weight_quantizers.192 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.gate_up_proj_weight_quantizers.193 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.gate_up_proj_weight_quantizers.194 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.gate_up_proj_weight_quantizers.195 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.84e-01 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.gate_up_proj_weight_quantizers.196 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.gate_up_proj_weight_quantizers.197 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.gate_up_proj_weight_quantizers.198 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=4.08e-01 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.gate_up_proj_weight_quantizers.199 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.gate_up_proj_weight_quantizers.200 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.28e-01 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.gate_up_proj_weight_quantizers.201 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.gate_up_proj_weight_quantizers.202 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.gate_up_proj_weight_quantizers.203 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.73e-01 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.gate_up_proj_weight_quantizers.204 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.gate_up_proj_weight_quantizers.205 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.gate_up_proj_weight_quantizers.206 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.gate_up_proj_weight_quantizers.207 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.gate_up_proj_weight_quantizers.208 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.gate_up_proj_weight_quantizers.209 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.55e-01 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.gate_up_proj_weight_quantizers.210 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.gate_up_proj_weight_quantizers.211 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.gate_up_proj_weight_quantizers.212 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.38e-01 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.gate_up_proj_weight_quantizers.213 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.15e-01 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.gate_up_proj_weight_quantizers.214 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.gate_up_proj_weight_quantizers.215 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.82e-01 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.gate_up_proj_weight_quantizers.216 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.24e-01 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.gate_up_proj_weight_quantizers.217 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.gate_up_proj_weight_quantizers.218 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.19e-01 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.gate_up_proj_weight_quantizers.219 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.50e-01 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.gate_up_proj_weight_quantizers.220 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.25e-01 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.gate_up_proj_weight_quantizers.221 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.gate_up_proj_weight_quantizers.222 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.gate_up_proj_weight_quantizers.223 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.06e-01 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.gate_up_proj_weight_quantizers.224 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.gate_up_proj_weight_quantizers.225 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.gate_up_proj_weight_quantizers.226 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.gate_up_proj_weight_quantizers.227 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.gate_up_proj_weight_quantizers.228 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.gate_up_proj_weight_quantizers.229 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.20e-01 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.gate_up_proj_weight_quantizers.230 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.gate_up_proj_weight_quantizers.231 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.51e-01 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.gate_up_proj_weight_quantizers.232 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.gate_up_proj_weight_quantizers.233 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.56e-01 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.gate_up_proj_weight_quantizers.234 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.gate_up_proj_weight_quantizers.235 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.19e-01 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.gate_up_proj_weight_quantizers.236 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.15e-01 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.gate_up_proj_weight_quantizers.237 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.gate_up_proj_weight_quantizers.238 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.31e-01 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.gate_up_proj_weight_quantizers.239 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.gate_up_proj_weight_quantizers.240 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.gate_up_proj_weight_quantizers.241 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.21e-01 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.gate_up_proj_weight_quantizers.242 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.gate_up_proj_weight_quantizers.243 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.78e-01 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.gate_up_proj_weight_quantizers.244 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.42e-01 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.gate_up_proj_weight_quantizers.245 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.gate_up_proj_weight_quantizers.246 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.14e-01 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.gate_up_proj_weight_quantizers.247 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.gate_up_proj_weight_quantizers.248 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.21e-01 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.gate_up_proj_weight_quantizers.249 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.gate_up_proj_weight_quantizers.250 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.gate_up_proj_weight_quantizers.251 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.27e-01 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.gate_up_proj_weight_quantizers.252 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.gate_up_proj_weight_quantizers.253 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.gate_up_proj_weight_quantizers.254 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=4.90e-01 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.gate_up_proj_weight_quantizers.255 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.down_proj_input_quantizer TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.30e+01 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.down_proj_weight_quantizers.0 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.down_proj_weight_quantizers.1 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.down_proj_weight_quantizers.2 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.down_proj_weight_quantizers.3 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.19e-01 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.down_proj_weight_quantizers.4 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.down_proj_weight_quantizers.5 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.down_proj_weight_quantizers.6 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.down_proj_weight_quantizers.7 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.41e-01 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.down_proj_weight_quantizers.8 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.down_proj_weight_quantizers.9 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.15e-01 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.down_proj_weight_quantizers.10 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.down_proj_weight_quantizers.11 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.down_proj_weight_quantizers.12 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.89e-01 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.down_proj_weight_quantizers.13 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.down_proj_weight_quantizers.14 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.down_proj_weight_quantizers.15 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.down_proj_weight_quantizers.16 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.down_proj_weight_quantizers.17 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.78e-01 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.down_proj_weight_quantizers.18 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.down_proj_weight_quantizers.19 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.down_proj_weight_quantizers.20 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.down_proj_weight_quantizers.21 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.41e-01 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.down_proj_weight_quantizers.22 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.down_proj_weight_quantizers.23 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.53e-01 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.down_proj_weight_quantizers.24 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.down_proj_weight_quantizers.25 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.down_proj_weight_quantizers.26 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.down_proj_weight_quantizers.27 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.38e-01 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.down_proj_weight_quantizers.28 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.down_proj_weight_quantizers.29 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.down_proj_weight_quantizers.30 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.down_proj_weight_quantizers.31 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.down_proj_weight_quantizers.32 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.down_proj_weight_quantizers.33 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.down_proj_weight_quantizers.34 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.25e-01 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.down_proj_weight_quantizers.35 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.down_proj_weight_quantizers.36 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.30e-01 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.down_proj_weight_quantizers.37 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.down_proj_weight_quantizers.38 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.down_proj_weight_quantizers.39 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.down_proj_weight_quantizers.40 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.down_proj_weight_quantizers.41 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.down_proj_weight_quantizers.42 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.39e-01 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.down_proj_weight_quantizers.43 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.72e-01 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.down_proj_weight_quantizers.44 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.down_proj_weight_quantizers.45 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.64e-01 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.down_proj_weight_quantizers.46 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.down_proj_weight_quantizers.47 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.down_proj_weight_quantizers.48 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.38e-01 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.down_proj_weight_quantizers.49 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.down_proj_weight_quantizers.50 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.down_proj_weight_quantizers.51 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.25e-01 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.down_proj_weight_quantizers.52 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.down_proj_weight_quantizers.53 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.57e-01 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.down_proj_weight_quantizers.54 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.91e-01 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.down_proj_weight_quantizers.55 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.down_proj_weight_quantizers.56 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.down_proj_weight_quantizers.57 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.down_proj_weight_quantizers.58 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.93e-01 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.down_proj_weight_quantizers.59 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.down_proj_weight_quantizers.60 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.down_proj_weight_quantizers.61 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.29e-01 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.down_proj_weight_quantizers.62 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.down_proj_weight_quantizers.63 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.64e-02 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.down_proj_weight_quantizers.64 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.down_proj_weight_quantizers.65 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.down_proj_weight_quantizers.66 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.down_proj_weight_quantizers.67 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.down_proj_weight_quantizers.68 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.down_proj_weight_quantizers.69 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.down_proj_weight_quantizers.70 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.down_proj_weight_quantizers.71 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.50e-01 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.down_proj_weight_quantizers.72 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.down_proj_weight_quantizers.73 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.down_proj_weight_quantizers.74 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.down_proj_weight_quantizers.75 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.down_proj_weight_quantizers.76 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.down_proj_weight_quantizers.77 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.down_proj_weight_quantizers.78 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.15e-01 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.down_proj_weight_quantizers.79 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.down_proj_weight_quantizers.80 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.down_proj_weight_quantizers.81 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.down_proj_weight_quantizers.82 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.15e-01 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.down_proj_weight_quantizers.83 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.down_proj_weight_quantizers.84 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.77e-01 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.down_proj_weight_quantizers.85 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.down_proj_weight_quantizers.86 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.28e-01 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.down_proj_weight_quantizers.87 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.43e-01 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.down_proj_weight_quantizers.88 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.down_proj_weight_quantizers.89 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.48e-01 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.down_proj_weight_quantizers.90 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.down_proj_weight_quantizers.91 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.down_proj_weight_quantizers.92 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.down_proj_weight_quantizers.93 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.down_proj_weight_quantizers.94 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.down_proj_weight_quantizers.95 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.90e-01 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.down_proj_weight_quantizers.96 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.down_proj_weight_quantizers.97 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.down_proj_weight_quantizers.98 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.down_proj_weight_quantizers.99 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.29e-01 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.down_proj_weight_quantizers.100 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.down_proj_weight_quantizers.101 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.down_proj_weight_quantizers.102 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.69e-01 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.down_proj_weight_quantizers.103 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.down_proj_weight_quantizers.104 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.down_proj_weight_quantizers.105 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.down_proj_weight_quantizers.106 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.down_proj_weight_quantizers.107 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.20e-01 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.down_proj_weight_quantizers.108 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.down_proj_weight_quantizers.109 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.down_proj_weight_quantizers.110 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.64e-01 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.down_proj_weight_quantizers.111 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.86e-01 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.down_proj_weight_quantizers.112 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.down_proj_weight_quantizers.113 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.down_proj_weight_quantizers.114 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.down_proj_weight_quantizers.115 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.91e-01 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.down_proj_weight_quantizers.116 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.down_proj_weight_quantizers.117 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.down_proj_weight_quantizers.118 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.down_proj_weight_quantizers.119 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.down_proj_weight_quantizers.120 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.down_proj_weight_quantizers.121 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.down_proj_weight_quantizers.122 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.down_proj_weight_quantizers.123 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.down_proj_weight_quantizers.124 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.down_proj_weight_quantizers.125 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.down_proj_weight_quantizers.126 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.down_proj_weight_quantizers.127 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.down_proj_weight_quantizers.128 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.down_proj_weight_quantizers.129 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.down_proj_weight_quantizers.130 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.down_proj_weight_quantizers.131 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.down_proj_weight_quantizers.132 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.down_proj_weight_quantizers.133 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.down_proj_weight_quantizers.134 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.down_proj_weight_quantizers.135 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.16e-01 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.down_proj_weight_quantizers.136 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.down_proj_weight_quantizers.137 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.down_proj_weight_quantizers.138 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.down_proj_weight_quantizers.139 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.62e-01 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.down_proj_weight_quantizers.140 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.down_proj_weight_quantizers.141 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.down_proj_weight_quantizers.142 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.58e-01 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.down_proj_weight_quantizers.143 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.down_proj_weight_quantizers.144 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.down_proj_weight_quantizers.145 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.down_proj_weight_quantizers.146 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.down_proj_weight_quantizers.147 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.down_proj_weight_quantizers.148 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.down_proj_weight_quantizers.149 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.down_proj_weight_quantizers.150 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.down_proj_weight_quantizers.151 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.down_proj_weight_quantizers.152 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.down_proj_weight_quantizers.153 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.16e-01 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.down_proj_weight_quantizers.154 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.down_proj_weight_quantizers.155 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.41e-01 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.down_proj_weight_quantizers.156 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.56e-01 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.down_proj_weight_quantizers.157 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.down_proj_weight_quantizers.158 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.down_proj_weight_quantizers.159 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.down_proj_weight_quantizers.160 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.down_proj_weight_quantizers.161 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.down_proj_weight_quantizers.162 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.down_proj_weight_quantizers.163 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.down_proj_weight_quantizers.164 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.down_proj_weight_quantizers.165 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.down_proj_weight_quantizers.166 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.down_proj_weight_quantizers.167 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.61e-01 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.down_proj_weight_quantizers.168 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.down_proj_weight_quantizers.169 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.down_proj_weight_quantizers.170 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.down_proj_weight_quantizers.171 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.47e-01 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.down_proj_weight_quantizers.172 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.17e-01 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.down_proj_weight_quantizers.173 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.down_proj_weight_quantizers.174 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.down_proj_weight_quantizers.175 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.down_proj_weight_quantizers.176 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.27e-01 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.down_proj_weight_quantizers.177 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.down_proj_weight_quantizers.178 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.42e-01 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.down_proj_weight_quantizers.179 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.down_proj_weight_quantizers.180 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.18e-01 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.down_proj_weight_quantizers.181 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.down_proj_weight_quantizers.182 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.down_proj_weight_quantizers.183 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.down_proj_weight_quantizers.184 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.down_proj_weight_quantizers.185 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.down_proj_weight_quantizers.186 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.down_proj_weight_quantizers.187 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.down_proj_weight_quantizers.188 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.down_proj_weight_quantizers.189 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.down_proj_weight_quantizers.190 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.down_proj_weight_quantizers.191 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.down_proj_weight_quantizers.192 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.down_proj_weight_quantizers.193 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.down_proj_weight_quantizers.194 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.down_proj_weight_quantizers.195 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.25e-01 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.down_proj_weight_quantizers.196 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.down_proj_weight_quantizers.197 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.down_proj_weight_quantizers.198 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.77e-01 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.down_proj_weight_quantizers.199 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.down_proj_weight_quantizers.200 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.31e-01 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.down_proj_weight_quantizers.201 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.down_proj_weight_quantizers.202 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.down_proj_weight_quantizers.203 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.38e-01 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.down_proj_weight_quantizers.204 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.down_proj_weight_quantizers.205 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.down_proj_weight_quantizers.206 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.down_proj_weight_quantizers.207 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.down_proj_weight_quantizers.208 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.down_proj_weight_quantizers.209 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.29e-01 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.down_proj_weight_quantizers.210 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.down_proj_weight_quantizers.211 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.down_proj_weight_quantizers.212 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.72e-01 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.down_proj_weight_quantizers.213 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.down_proj_weight_quantizers.214 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.down_proj_weight_quantizers.215 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.69e-01 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.down_proj_weight_quantizers.216 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.down_proj_weight_quantizers.217 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.down_proj_weight_quantizers.218 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.down_proj_weight_quantizers.219 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.50e-01 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.down_proj_weight_quantizers.220 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.47e-01 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.down_proj_weight_quantizers.221 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.down_proj_weight_quantizers.222 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.down_proj_weight_quantizers.223 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.down_proj_weight_quantizers.224 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.down_proj_weight_quantizers.225 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.19e-01 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.down_proj_weight_quantizers.226 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.13e-01 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.down_proj_weight_quantizers.227 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.down_proj_weight_quantizers.228 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.down_proj_weight_quantizers.229 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.99e-01 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.down_proj_weight_quantizers.230 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.down_proj_weight_quantizers.231 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.down_proj_weight_quantizers.232 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.down_proj_weight_quantizers.233 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.down_proj_weight_quantizers.234 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.down_proj_weight_quantizers.235 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.down_proj_weight_quantizers.236 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.18e-01 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.down_proj_weight_quantizers.237 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.down_proj_weight_quantizers.238 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.68e-01 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.down_proj_weight_quantizers.239 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.down_proj_weight_quantizers.240 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.down_proj_weight_quantizers.241 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.down_proj_weight_quantizers.242 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.64e-02 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.down_proj_weight_quantizers.243 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.58e-01 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.down_proj_weight_quantizers.244 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.down_proj_weight_quantizers.245 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.down_proj_weight_quantizers.246 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.down_proj_weight_quantizers.247 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.down_proj_weight_quantizers.248 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.down_proj_weight_quantizers.249 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.down_proj_weight_quantizers.250 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.down_proj_weight_quantizers.251 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.down_proj_weight_quantizers.252 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.down_proj_weight_quantizers.253 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.down_proj_weight_quantizers.254 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.93e-01 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.down_proj_weight_quantizers.255 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.68.mlp.shared_experts.gate_proj.input_quantizer TensorQuantizer(disabled) +model.layers.68.mlp.shared_experts.gate_proj.output_quantizer TensorQuantizer(disabled) +model.layers.68.mlp.shared_experts.gate_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.68.mlp.shared_experts.up_proj.input_quantizer TensorQuantizer(disabled) +model.layers.68.mlp.shared_experts.up_proj.output_quantizer TensorQuantizer(disabled) +model.layers.68.mlp.shared_experts.up_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.68.mlp.shared_experts.down_proj.input_quantizer TensorQuantizer(disabled) +model.layers.68.mlp.shared_experts.down_proj.output_quantizer TensorQuantizer(disabled) +model.layers.68.mlp.shared_experts.down_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.69.self_attn.q_a_proj.input_quantizer TensorQuantizer(disabled) +model.layers.69.self_attn.q_a_proj.output_quantizer TensorQuantizer(disabled) +model.layers.69.self_attn.q_a_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.69.self_attn.q_b_proj.input_quantizer TensorQuantizer(disabled) +model.layers.69.self_attn.q_b_proj.output_quantizer TensorQuantizer(disabled) +model.layers.69.self_attn.q_b_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.69.self_attn.kv_a_proj_with_mqa.input_quantizer TensorQuantizer(disabled) +model.layers.69.self_attn.kv_a_proj_with_mqa.output_quantizer TensorQuantizer(disabled) +model.layers.69.self_attn.kv_a_proj_with_mqa.weight_quantizer TensorQuantizer(disabled) +model.layers.69.self_attn.kv_b_proj.input_quantizer TensorQuantizer(disabled) +model.layers.69.self_attn.kv_b_proj.output_quantizer TensorQuantizer(disabled) +model.layers.69.self_attn.kv_b_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.69.self_attn.o_proj.input_quantizer TensorQuantizer(disabled) +model.layers.69.self_attn.o_proj.output_quantizer TensorQuantizer(disabled) +model.layers.69.self_attn.o_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.69.self_attn.q_bmm_quantizer TensorQuantizer(disabled) +model.layers.69.self_attn.k_bmm_quantizer TensorQuantizer((4, 3) bit fake per-tensor amax=4.48e+02(const) calibrator=MaxCalibrator quant) +model.layers.69.self_attn.v_bmm_quantizer TensorQuantizer((4, 3) bit fake per-tensor amax=4.48e+02(const) calibrator=MaxCalibrator quant) +model.layers.69.self_attn.p_bmm_quantizer TensorQuantizer(disabled) +model.layers.69.mlp.experts.gate_up_proj_input_quantizer TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.28e+00 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.gate_up_proj_weight_quantizers.0 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.gate_up_proj_weight_quantizers.1 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.61e-01 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.gate_up_proj_weight_quantizers.2 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.gate_up_proj_weight_quantizers.3 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.gate_up_proj_weight_quantizers.4 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.17e-01 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.gate_up_proj_weight_quantizers.5 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.15e-01 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.gate_up_proj_weight_quantizers.6 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.54e-01 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.gate_up_proj_weight_quantizers.7 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.gate_up_proj_weight_quantizers.8 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.gate_up_proj_weight_quantizers.9 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.gate_up_proj_weight_quantizers.10 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.95e-01 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.gate_up_proj_weight_quantizers.11 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.87e-01 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.gate_up_proj_weight_quantizers.12 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.gate_up_proj_weight_quantizers.13 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.23e-01 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.gate_up_proj_weight_quantizers.14 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.gate_up_proj_weight_quantizers.15 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.gate_up_proj_weight_quantizers.16 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.gate_up_proj_weight_quantizers.17 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.22e-01 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.gate_up_proj_weight_quantizers.18 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.26e-01 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.gate_up_proj_weight_quantizers.19 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.25e-01 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.gate_up_proj_weight_quantizers.20 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.gate_up_proj_weight_quantizers.21 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.15e-01 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.gate_up_proj_weight_quantizers.22 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.18e-01 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.gate_up_proj_weight_quantizers.23 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.13e-01 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.gate_up_proj_weight_quantizers.24 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=5.20e-01 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.gate_up_proj_weight_quantizers.25 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.50e-01 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.gate_up_proj_weight_quantizers.26 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.71e-01 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.gate_up_proj_weight_quantizers.27 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.36e-01 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.gate_up_proj_weight_quantizers.28 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.gate_up_proj_weight_quantizers.29 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.46e-01 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.gate_up_proj_weight_quantizers.30 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.gate_up_proj_weight_quantizers.31 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.gate_up_proj_weight_quantizers.32 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.gate_up_proj_weight_quantizers.33 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.gate_up_proj_weight_quantizers.34 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.17e-01 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.gate_up_proj_weight_quantizers.35 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.gate_up_proj_weight_quantizers.36 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.73e-01 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.gate_up_proj_weight_quantizers.37 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.25e-01 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.gate_up_proj_weight_quantizers.38 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.gate_up_proj_weight_quantizers.39 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.gate_up_proj_weight_quantizers.40 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.15e-01 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.gate_up_proj_weight_quantizers.41 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.27e-01 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.gate_up_proj_weight_quantizers.42 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.23e-01 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.gate_up_proj_weight_quantizers.43 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.19e-01 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.gate_up_proj_weight_quantizers.44 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.28e-01 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.gate_up_proj_weight_quantizers.45 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.93e-01 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.gate_up_proj_weight_quantizers.46 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.gate_up_proj_weight_quantizers.47 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=6.64e-01 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.gate_up_proj_weight_quantizers.48 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.gate_up_proj_weight_quantizers.49 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.gate_up_proj_weight_quantizers.50 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.gate_up_proj_weight_quantizers.51 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.gate_up_proj_weight_quantizers.52 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.gate_up_proj_weight_quantizers.53 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.25e-01 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.gate_up_proj_weight_quantizers.54 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.gate_up_proj_weight_quantizers.55 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.gate_up_proj_weight_quantizers.56 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.gate_up_proj_weight_quantizers.57 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.gate_up_proj_weight_quantizers.58 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.48e-01 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.gate_up_proj_weight_quantizers.59 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.gate_up_proj_weight_quantizers.60 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.gate_up_proj_weight_quantizers.61 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.80e-01 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.gate_up_proj_weight_quantizers.62 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.gate_up_proj_weight_quantizers.63 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.24e-01 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.gate_up_proj_weight_quantizers.64 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.30e-01 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.gate_up_proj_weight_quantizers.65 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.gate_up_proj_weight_quantizers.66 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.gate_up_proj_weight_quantizers.67 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.gate_up_proj_weight_quantizers.68 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.gate_up_proj_weight_quantizers.69 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.gate_up_proj_weight_quantizers.70 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.55e-01 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.gate_up_proj_weight_quantizers.71 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.28e-01 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.gate_up_proj_weight_quantizers.72 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.15e-01 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.gate_up_proj_weight_quantizers.73 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.36e-01 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.gate_up_proj_weight_quantizers.74 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.gate_up_proj_weight_quantizers.75 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.gate_up_proj_weight_quantizers.76 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.50e-01 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.gate_up_proj_weight_quantizers.77 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.gate_up_proj_weight_quantizers.78 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.17e-01 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.gate_up_proj_weight_quantizers.79 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.28e-01 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.gate_up_proj_weight_quantizers.80 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.gate_up_proj_weight_quantizers.81 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.gate_up_proj_weight_quantizers.82 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.gate_up_proj_weight_quantizers.83 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.gate_up_proj_weight_quantizers.84 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.50e-01 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.gate_up_proj_weight_quantizers.85 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.17e-01 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.gate_up_proj_weight_quantizers.86 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.gate_up_proj_weight_quantizers.87 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=7.38e-01 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.gate_up_proj_weight_quantizers.88 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.15e-01 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.gate_up_proj_weight_quantizers.89 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.50e-01 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.gate_up_proj_weight_quantizers.90 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.gate_up_proj_weight_quantizers.91 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.gate_up_proj_weight_quantizers.92 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.gate_up_proj_weight_quantizers.93 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.18e-01 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.gate_up_proj_weight_quantizers.94 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.gate_up_proj_weight_quantizers.95 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.52e-01 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.gate_up_proj_weight_quantizers.96 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.82e-01 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.gate_up_proj_weight_quantizers.97 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.22e-01 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.gate_up_proj_weight_quantizers.98 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.gate_up_proj_weight_quantizers.99 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.gate_up_proj_weight_quantizers.100 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.27e-01 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.gate_up_proj_weight_quantizers.101 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.gate_up_proj_weight_quantizers.102 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.81e-01 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.gate_up_proj_weight_quantizers.103 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.15e-01 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.gate_up_proj_weight_quantizers.104 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.21e-01 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.gate_up_proj_weight_quantizers.105 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.gate_up_proj_weight_quantizers.106 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.29e-01 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.gate_up_proj_weight_quantizers.107 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.15e-01 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.gate_up_proj_weight_quantizers.108 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.26e-01 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.gate_up_proj_weight_quantizers.109 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.29e-01 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.gate_up_proj_weight_quantizers.110 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.gate_up_proj_weight_quantizers.111 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.gate_up_proj_weight_quantizers.112 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.gate_up_proj_weight_quantizers.113 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.gate_up_proj_weight_quantizers.114 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.55e-01 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.gate_up_proj_weight_quantizers.115 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.05e-01 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.gate_up_proj_weight_quantizers.116 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.42e-01 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.gate_up_proj_weight_quantizers.117 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.13e-01 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.gate_up_proj_weight_quantizers.118 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.gate_up_proj_weight_quantizers.119 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.gate_up_proj_weight_quantizers.120 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.70e-01 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.gate_up_proj_weight_quantizers.121 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.gate_up_proj_weight_quantizers.122 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.gate_up_proj_weight_quantizers.123 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.gate_up_proj_weight_quantizers.124 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.18e-01 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.gate_up_proj_weight_quantizers.125 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.58e-01 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.gate_up_proj_weight_quantizers.126 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.15e-01 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.gate_up_proj_weight_quantizers.127 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.gate_up_proj_weight_quantizers.128 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.gate_up_proj_weight_quantizers.129 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.46e-01 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.gate_up_proj_weight_quantizers.130 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.gate_up_proj_weight_quantizers.131 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.gate_up_proj_weight_quantizers.132 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.23e-01 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.gate_up_proj_weight_quantizers.133 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.gate_up_proj_weight_quantizers.134 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.76e-01 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.gate_up_proj_weight_quantizers.135 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.15e-01 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.gate_up_proj_weight_quantizers.136 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.gate_up_proj_weight_quantizers.137 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.26e-01 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.gate_up_proj_weight_quantizers.138 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.54e-01 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.gate_up_proj_weight_quantizers.139 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.gate_up_proj_weight_quantizers.140 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.gate_up_proj_weight_quantizers.141 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.23e-01 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.gate_up_proj_weight_quantizers.142 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.58e-01 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.gate_up_proj_weight_quantizers.143 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.15e-01 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.gate_up_proj_weight_quantizers.144 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.gate_up_proj_weight_quantizers.145 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.13e-01 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.gate_up_proj_weight_quantizers.146 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.gate_up_proj_weight_quantizers.147 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.25e-01 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.gate_up_proj_weight_quantizers.148 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.24e-01 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.gate_up_proj_weight_quantizers.149 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.55e-01 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.gate_up_proj_weight_quantizers.150 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.16e-01 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.gate_up_proj_weight_quantizers.151 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.17e-01 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.gate_up_proj_weight_quantizers.152 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.32e-01 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.gate_up_proj_weight_quantizers.153 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.gate_up_proj_weight_quantizers.154 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.gate_up_proj_weight_quantizers.155 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.17e-01 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.gate_up_proj_weight_quantizers.156 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.19e-01 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.gate_up_proj_weight_quantizers.157 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.31e-01 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.gate_up_proj_weight_quantizers.158 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.gate_up_proj_weight_quantizers.159 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.19e-01 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.gate_up_proj_weight_quantizers.160 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.gate_up_proj_weight_quantizers.161 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=6.48e-01 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.gate_up_proj_weight_quantizers.162 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.gate_up_proj_weight_quantizers.163 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.47e-01 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.gate_up_proj_weight_quantizers.164 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.gate_up_proj_weight_quantizers.165 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.gate_up_proj_weight_quantizers.166 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.56e-01 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.gate_up_proj_weight_quantizers.167 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.21e-01 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.gate_up_proj_weight_quantizers.168 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.23e-01 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.gate_up_proj_weight_quantizers.169 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.21e-01 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.gate_up_proj_weight_quantizers.170 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.gate_up_proj_weight_quantizers.171 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.48e-01 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.gate_up_proj_weight_quantizers.172 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.gate_up_proj_weight_quantizers.173 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.gate_up_proj_weight_quantizers.174 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.gate_up_proj_weight_quantizers.175 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.17e-01 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.gate_up_proj_weight_quantizers.176 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.56e-01 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.gate_up_proj_weight_quantizers.177 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.55e-01 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.gate_up_proj_weight_quantizers.178 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.gate_up_proj_weight_quantizers.179 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.gate_up_proj_weight_quantizers.180 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.22e-01 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.gate_up_proj_weight_quantizers.181 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.17e-01 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.gate_up_proj_weight_quantizers.182 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.gate_up_proj_weight_quantizers.183 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.gate_up_proj_weight_quantizers.184 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.37e-01 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.gate_up_proj_weight_quantizers.185 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.31e-01 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.gate_up_proj_weight_quantizers.186 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.gate_up_proj_weight_quantizers.187 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.34e-01 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.gate_up_proj_weight_quantizers.188 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.gate_up_proj_weight_quantizers.189 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.21e-01 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.gate_up_proj_weight_quantizers.190 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.gate_up_proj_weight_quantizers.191 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.gate_up_proj_weight_quantizers.192 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.gate_up_proj_weight_quantizers.193 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.19e-01 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.gate_up_proj_weight_quantizers.194 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.gate_up_proj_weight_quantizers.195 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.gate_up_proj_weight_quantizers.196 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.gate_up_proj_weight_quantizers.197 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.47e-01 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.gate_up_proj_weight_quantizers.198 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.95e-01 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.gate_up_proj_weight_quantizers.199 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.19e-01 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.gate_up_proj_weight_quantizers.200 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.22e-01 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.gate_up_proj_weight_quantizers.201 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.gate_up_proj_weight_quantizers.202 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.gate_up_proj_weight_quantizers.203 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.gate_up_proj_weight_quantizers.204 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.gate_up_proj_weight_quantizers.205 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.gate_up_proj_weight_quantizers.206 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.20e-01 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.gate_up_proj_weight_quantizers.207 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.gate_up_proj_weight_quantizers.208 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.gate_up_proj_weight_quantizers.209 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.29e-01 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.gate_up_proj_weight_quantizers.210 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.24e-01 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.gate_up_proj_weight_quantizers.211 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.19e-01 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.gate_up_proj_weight_quantizers.212 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.97e-01 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.gate_up_proj_weight_quantizers.213 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.gate_up_proj_weight_quantizers.214 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.gate_up_proj_weight_quantizers.215 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.73e-01 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.gate_up_proj_weight_quantizers.216 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.gate_up_proj_weight_quantizers.217 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.21e-01 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.gate_up_proj_weight_quantizers.218 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.gate_up_proj_weight_quantizers.219 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.gate_up_proj_weight_quantizers.220 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.gate_up_proj_weight_quantizers.221 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.89e-01 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.gate_up_proj_weight_quantizers.222 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.31e-01 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.gate_up_proj_weight_quantizers.223 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.gate_up_proj_weight_quantizers.224 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.42e-01 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.gate_up_proj_weight_quantizers.225 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.gate_up_proj_weight_quantizers.226 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.59e-01 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.gate_up_proj_weight_quantizers.227 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.32e-01 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.gate_up_proj_weight_quantizers.228 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.52e-01 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.gate_up_proj_weight_quantizers.229 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.gate_up_proj_weight_quantizers.230 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.53e-01 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.gate_up_proj_weight_quantizers.231 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.24e-01 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.gate_up_proj_weight_quantizers.232 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.gate_up_proj_weight_quantizers.233 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.gate_up_proj_weight_quantizers.234 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.gate_up_proj_weight_quantizers.235 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-01 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.gate_up_proj_weight_quantizers.236 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.gate_up_proj_weight_quantizers.237 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.13e-01 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.gate_up_proj_weight_quantizers.238 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.gate_up_proj_weight_quantizers.239 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.gate_up_proj_weight_quantizers.240 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.gate_up_proj_weight_quantizers.241 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.28e-01 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.gate_up_proj_weight_quantizers.242 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.gate_up_proj_weight_quantizers.243 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.gate_up_proj_weight_quantizers.244 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.81e-01 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.gate_up_proj_weight_quantizers.245 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.91e-01 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.gate_up_proj_weight_quantizers.246 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.71e-01 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.gate_up_proj_weight_quantizers.247 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.gate_up_proj_weight_quantizers.248 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.gate_up_proj_weight_quantizers.249 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.gate_up_proj_weight_quantizers.250 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.gate_up_proj_weight_quantizers.251 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.gate_up_proj_weight_quantizers.252 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.20e-01 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.gate_up_proj_weight_quantizers.253 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.05e-01 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.gate_up_proj_weight_quantizers.254 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.24e-01 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.gate_up_proj_weight_quantizers.255 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.down_proj_input_quantizer TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=7.10e+01 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.down_proj_weight_quantizers.0 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.down_proj_weight_quantizers.1 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.down_proj_weight_quantizers.2 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.down_proj_weight_quantizers.3 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.down_proj_weight_quantizers.4 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.down_proj_weight_quantizers.5 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.down_proj_weight_quantizers.6 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.43e-01 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.down_proj_weight_quantizers.7 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.down_proj_weight_quantizers.8 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.down_proj_weight_quantizers.9 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.48e-01 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.down_proj_weight_quantizers.10 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.32e-01 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.down_proj_weight_quantizers.11 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.60e-01 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.down_proj_weight_quantizers.12 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.down_proj_weight_quantizers.13 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.31e-01 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.down_proj_weight_quantizers.14 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.down_proj_weight_quantizers.15 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.down_proj_weight_quantizers.16 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.down_proj_weight_quantizers.17 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.down_proj_weight_quantizers.18 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.39e-01 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.down_proj_weight_quantizers.19 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.down_proj_weight_quantizers.20 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.down_proj_weight_quantizers.21 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.down_proj_weight_quantizers.22 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.down_proj_weight_quantizers.23 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.down_proj_weight_quantizers.24 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.03e-01 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.down_proj_weight_quantizers.25 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.down_proj_weight_quantizers.26 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.down_proj_weight_quantizers.27 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.down_proj_weight_quantizers.28 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.down_proj_weight_quantizers.29 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.down_proj_weight_quantizers.30 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.down_proj_weight_quantizers.31 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.down_proj_weight_quantizers.32 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.down_proj_weight_quantizers.33 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.down_proj_weight_quantizers.34 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.down_proj_weight_quantizers.35 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.down_proj_weight_quantizers.36 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.17e-01 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.down_proj_weight_quantizers.37 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.down_proj_weight_quantizers.38 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.down_proj_weight_quantizers.39 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.down_proj_weight_quantizers.40 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.down_proj_weight_quantizers.41 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.down_proj_weight_quantizers.42 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.down_proj_weight_quantizers.43 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.down_proj_weight_quantizers.44 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.32e-01 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.down_proj_weight_quantizers.45 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.25e-01 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.down_proj_weight_quantizers.46 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.down_proj_weight_quantizers.47 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.72e-01 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.down_proj_weight_quantizers.48 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.down_proj_weight_quantizers.49 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.down_proj_weight_quantizers.50 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.49e-01 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.down_proj_weight_quantizers.51 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.down_proj_weight_quantizers.52 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.down_proj_weight_quantizers.53 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.down_proj_weight_quantizers.54 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.down_proj_weight_quantizers.55 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.down_proj_weight_quantizers.56 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.down_proj_weight_quantizers.57 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.down_proj_weight_quantizers.58 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.down_proj_weight_quantizers.59 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.down_proj_weight_quantizers.60 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.down_proj_weight_quantizers.61 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.down_proj_weight_quantizers.62 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.down_proj_weight_quantizers.63 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.down_proj_weight_quantizers.64 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.down_proj_weight_quantizers.65 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.down_proj_weight_quantizers.66 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.down_proj_weight_quantizers.67 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.down_proj_weight_quantizers.68 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.down_proj_weight_quantizers.69 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.down_proj_weight_quantizers.70 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.75e-01 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.down_proj_weight_quantizers.71 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.24e-01 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.down_proj_weight_quantizers.72 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.down_proj_weight_quantizers.73 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.down_proj_weight_quantizers.74 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.down_proj_weight_quantizers.75 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.down_proj_weight_quantizers.76 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.94e-01 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.down_proj_weight_quantizers.77 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.down_proj_weight_quantizers.78 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.14e-01 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.down_proj_weight_quantizers.79 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.down_proj_weight_quantizers.80 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.down_proj_weight_quantizers.81 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.down_proj_weight_quantizers.82 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.down_proj_weight_quantizers.83 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.down_proj_weight_quantizers.84 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.down_proj_weight_quantizers.85 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.down_proj_weight_quantizers.86 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.down_proj_weight_quantizers.87 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.50e-01 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.down_proj_weight_quantizers.88 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.down_proj_weight_quantizers.89 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.down_proj_weight_quantizers.90 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.down_proj_weight_quantizers.91 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.down_proj_weight_quantizers.92 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.down_proj_weight_quantizers.93 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.down_proj_weight_quantizers.94 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.down_proj_weight_quantizers.95 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.23e-01 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.down_proj_weight_quantizers.96 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.67e-01 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.down_proj_weight_quantizers.97 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.down_proj_weight_quantizers.98 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.38e-01 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.down_proj_weight_quantizers.99 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.down_proj_weight_quantizers.100 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.down_proj_weight_quantizers.101 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.down_proj_weight_quantizers.102 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.down_proj_weight_quantizers.103 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.down_proj_weight_quantizers.104 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.32e-01 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.down_proj_weight_quantizers.105 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.down_proj_weight_quantizers.106 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.down_proj_weight_quantizers.107 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.17e-01 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.down_proj_weight_quantizers.108 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.down_proj_weight_quantizers.109 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.down_proj_weight_quantizers.110 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.down_proj_weight_quantizers.111 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.down_proj_weight_quantizers.112 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.down_proj_weight_quantizers.113 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.down_proj_weight_quantizers.114 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.down_proj_weight_quantizers.115 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.down_proj_weight_quantizers.116 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.down_proj_weight_quantizers.117 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.57e-01 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.down_proj_weight_quantizers.118 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.down_proj_weight_quantizers.119 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.down_proj_weight_quantizers.120 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.35e-01 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.down_proj_weight_quantizers.121 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.down_proj_weight_quantizers.122 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.down_proj_weight_quantizers.123 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.down_proj_weight_quantizers.124 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.down_proj_weight_quantizers.125 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.down_proj_weight_quantizers.126 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.16e-01 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.down_proj_weight_quantizers.127 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.15e-01 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.down_proj_weight_quantizers.128 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.down_proj_weight_quantizers.129 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.20e-01 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.down_proj_weight_quantizers.130 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.down_proj_weight_quantizers.131 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.down_proj_weight_quantizers.132 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.down_proj_weight_quantizers.133 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.down_proj_weight_quantizers.134 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.01e-01 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.down_proj_weight_quantizers.135 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.down_proj_weight_quantizers.136 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.down_proj_weight_quantizers.137 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.down_proj_weight_quantizers.138 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.down_proj_weight_quantizers.139 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.59e-02 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.down_proj_weight_quantizers.140 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.down_proj_weight_quantizers.141 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.down_proj_weight_quantizers.142 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.down_proj_weight_quantizers.143 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.down_proj_weight_quantizers.144 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.51e-01 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.down_proj_weight_quantizers.145 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.down_proj_weight_quantizers.146 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.down_proj_weight_quantizers.147 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.down_proj_weight_quantizers.148 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.down_proj_weight_quantizers.149 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.67e-01 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.down_proj_weight_quantizers.150 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.73e-01 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.down_proj_weight_quantizers.151 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.down_proj_weight_quantizers.152 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.down_proj_weight_quantizers.153 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.down_proj_weight_quantizers.154 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.17e-01 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.down_proj_weight_quantizers.155 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.down_proj_weight_quantizers.156 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.down_proj_weight_quantizers.157 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.down_proj_weight_quantizers.158 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.down_proj_weight_quantizers.159 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.down_proj_weight_quantizers.160 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.down_proj_weight_quantizers.161 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.03e-01 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.down_proj_weight_quantizers.162 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.down_proj_weight_quantizers.163 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.down_proj_weight_quantizers.164 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.down_proj_weight_quantizers.165 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.20e-01 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.down_proj_weight_quantizers.166 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.18e-01 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.down_proj_weight_quantizers.167 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.down_proj_weight_quantizers.168 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.down_proj_weight_quantizers.169 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.down_proj_weight_quantizers.170 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.down_proj_weight_quantizers.171 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.27e-01 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.down_proj_weight_quantizers.172 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.29e-01 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.down_proj_weight_quantizers.173 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.25e-01 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.down_proj_weight_quantizers.174 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.down_proj_weight_quantizers.175 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.down_proj_weight_quantizers.176 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.down_proj_weight_quantizers.177 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.15e-01 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.down_proj_weight_quantizers.178 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.down_proj_weight_quantizers.179 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.down_proj_weight_quantizers.180 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.down_proj_weight_quantizers.181 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.down_proj_weight_quantizers.182 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.down_proj_weight_quantizers.183 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.down_proj_weight_quantizers.184 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.35e-01 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.down_proj_weight_quantizers.185 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.down_proj_weight_quantizers.186 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.down_proj_weight_quantizers.187 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.16e-01 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.down_proj_weight_quantizers.188 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.down_proj_weight_quantizers.189 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.down_proj_weight_quantizers.190 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.down_proj_weight_quantizers.191 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.down_proj_weight_quantizers.192 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.down_proj_weight_quantizers.193 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.down_proj_weight_quantizers.194 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.down_proj_weight_quantizers.195 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.down_proj_weight_quantizers.196 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.down_proj_weight_quantizers.197 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.down_proj_weight_quantizers.198 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.18e-01 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.down_proj_weight_quantizers.199 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.down_proj_weight_quantizers.200 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.down_proj_weight_quantizers.201 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.down_proj_weight_quantizers.202 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.45e-01 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.down_proj_weight_quantizers.203 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.down_proj_weight_quantizers.204 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.down_proj_weight_quantizers.205 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.down_proj_weight_quantizers.206 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.down_proj_weight_quantizers.207 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.down_proj_weight_quantizers.208 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.down_proj_weight_quantizers.209 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.down_proj_weight_quantizers.210 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.down_proj_weight_quantizers.211 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.down_proj_weight_quantizers.212 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.18e-01 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.down_proj_weight_quantizers.213 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.down_proj_weight_quantizers.214 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.down_proj_weight_quantizers.215 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.37e-01 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.down_proj_weight_quantizers.216 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.down_proj_weight_quantizers.217 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.down_proj_weight_quantizers.218 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.49e-01 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.down_proj_weight_quantizers.219 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.32e-01 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.down_proj_weight_quantizers.220 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.down_proj_weight_quantizers.221 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.69e-01 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.down_proj_weight_quantizers.222 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.26e-01 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.down_proj_weight_quantizers.223 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.down_proj_weight_quantizers.224 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.58e-01 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.down_proj_weight_quantizers.225 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.down_proj_weight_quantizers.226 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.down_proj_weight_quantizers.227 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.down_proj_weight_quantizers.228 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.30e-01 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.down_proj_weight_quantizers.229 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.down_proj_weight_quantizers.230 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.down_proj_weight_quantizers.231 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.52e-01 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.down_proj_weight_quantizers.232 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.down_proj_weight_quantizers.233 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.down_proj_weight_quantizers.234 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.down_proj_weight_quantizers.235 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.16e-01 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.down_proj_weight_quantizers.236 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.down_proj_weight_quantizers.237 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.down_proj_weight_quantizers.238 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.down_proj_weight_quantizers.239 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.down_proj_weight_quantizers.240 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.down_proj_weight_quantizers.241 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.down_proj_weight_quantizers.242 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.16e-01 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.down_proj_weight_quantizers.243 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.down_proj_weight_quantizers.244 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.27e-01 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.down_proj_weight_quantizers.245 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.37e-01 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.down_proj_weight_quantizers.246 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.down_proj_weight_quantizers.247 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.down_proj_weight_quantizers.248 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.02e-01 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.down_proj_weight_quantizers.249 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.down_proj_weight_quantizers.250 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.down_proj_weight_quantizers.251 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.down_proj_weight_quantizers.252 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.down_proj_weight_quantizers.253 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.97e-01 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.down_proj_weight_quantizers.254 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.down_proj_weight_quantizers.255 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.69.mlp.shared_experts.gate_proj.input_quantizer TensorQuantizer(disabled) +model.layers.69.mlp.shared_experts.gate_proj.output_quantizer TensorQuantizer(disabled) +model.layers.69.mlp.shared_experts.gate_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.69.mlp.shared_experts.up_proj.input_quantizer TensorQuantizer(disabled) +model.layers.69.mlp.shared_experts.up_proj.output_quantizer TensorQuantizer(disabled) +model.layers.69.mlp.shared_experts.up_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.69.mlp.shared_experts.down_proj.input_quantizer TensorQuantizer(disabled) +model.layers.69.mlp.shared_experts.down_proj.output_quantizer TensorQuantizer(disabled) +model.layers.69.mlp.shared_experts.down_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.70.self_attn.q_a_proj.input_quantizer TensorQuantizer(disabled) +model.layers.70.self_attn.q_a_proj.output_quantizer TensorQuantizer(disabled) +model.layers.70.self_attn.q_a_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.70.self_attn.q_b_proj.input_quantizer TensorQuantizer(disabled) +model.layers.70.self_attn.q_b_proj.output_quantizer TensorQuantizer(disabled) +model.layers.70.self_attn.q_b_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.70.self_attn.kv_a_proj_with_mqa.input_quantizer TensorQuantizer(disabled) +model.layers.70.self_attn.kv_a_proj_with_mqa.output_quantizer TensorQuantizer(disabled) +model.layers.70.self_attn.kv_a_proj_with_mqa.weight_quantizer TensorQuantizer(disabled) +model.layers.70.self_attn.kv_b_proj.input_quantizer TensorQuantizer(disabled) +model.layers.70.self_attn.kv_b_proj.output_quantizer TensorQuantizer(disabled) +model.layers.70.self_attn.kv_b_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.70.self_attn.o_proj.input_quantizer TensorQuantizer(disabled) +model.layers.70.self_attn.o_proj.output_quantizer TensorQuantizer(disabled) +model.layers.70.self_attn.o_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.70.self_attn.indexer.wq_b.input_quantizer TensorQuantizer(disabled) +model.layers.70.self_attn.indexer.wq_b.output_quantizer TensorQuantizer(disabled) +model.layers.70.self_attn.indexer.wq_b.weight_quantizer TensorQuantizer(disabled) +model.layers.70.self_attn.indexer.wk.input_quantizer TensorQuantizer(disabled) +model.layers.70.self_attn.indexer.wk.output_quantizer TensorQuantizer(disabled) +model.layers.70.self_attn.indexer.wk.weight_quantizer TensorQuantizer(disabled) +model.layers.70.self_attn.indexer.k_norm.input_quantizer TensorQuantizer(disabled) +model.layers.70.self_attn.indexer.k_norm.output_quantizer TensorQuantizer(disabled) +model.layers.70.self_attn.indexer.weights_proj.input_quantizer TensorQuantizer(disabled) +model.layers.70.self_attn.indexer.weights_proj.output_quantizer TensorQuantizer(disabled) +model.layers.70.self_attn.indexer.weights_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.70.self_attn.q_bmm_quantizer TensorQuantizer(disabled) +model.layers.70.self_attn.k_bmm_quantizer TensorQuantizer((4, 3) bit fake per-tensor amax=4.48e+02(const) calibrator=MaxCalibrator quant) +model.layers.70.self_attn.v_bmm_quantizer TensorQuantizer((4, 3) bit fake per-tensor amax=4.48e+02(const) calibrator=MaxCalibrator quant) +model.layers.70.self_attn.p_bmm_quantizer TensorQuantizer(disabled) +model.layers.70.mlp.experts.gate_up_proj_input_quantizer TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.44e+00 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.gate_up_proj_weight_quantizers.0 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.35e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.gate_up_proj_weight_quantizers.1 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.05e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.gate_up_proj_weight_quantizers.2 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.gate_up_proj_weight_quantizers.3 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.gate_up_proj_weight_quantizers.4 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.96e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.gate_up_proj_weight_quantizers.5 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.gate_up_proj_weight_quantizers.6 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.38e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.gate_up_proj_weight_quantizers.7 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.76e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.gate_up_proj_weight_quantizers.8 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.gate_up_proj_weight_quantizers.9 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.44e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.gate_up_proj_weight_quantizers.10 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.17e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.gate_up_proj_weight_quantizers.11 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.19e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.gate_up_proj_weight_quantizers.12 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.80e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.gate_up_proj_weight_quantizers.13 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.gate_up_proj_weight_quantizers.14 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.gate_up_proj_weight_quantizers.15 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.gate_up_proj_weight_quantizers.16 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.18e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.gate_up_proj_weight_quantizers.17 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.19e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.gate_up_proj_weight_quantizers.18 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.17e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.gate_up_proj_weight_quantizers.19 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.23e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.gate_up_proj_weight_quantizers.20 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.65e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.gate_up_proj_weight_quantizers.21 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.gate_up_proj_weight_quantizers.22 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.16e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.gate_up_proj_weight_quantizers.23 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.59e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.gate_up_proj_weight_quantizers.24 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.gate_up_proj_weight_quantizers.25 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.gate_up_proj_weight_quantizers.26 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.gate_up_proj_weight_quantizers.27 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.gate_up_proj_weight_quantizers.28 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.16e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.gate_up_proj_weight_quantizers.29 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.28e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.gate_up_proj_weight_quantizers.30 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.75e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.gate_up_proj_weight_quantizers.31 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.gate_up_proj_weight_quantizers.32 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.95e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.gate_up_proj_weight_quantizers.33 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.gate_up_proj_weight_quantizers.34 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=4.84e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.gate_up_proj_weight_quantizers.35 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.39e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.gate_up_proj_weight_quantizers.36 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.18e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.gate_up_proj_weight_quantizers.37 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.65e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.gate_up_proj_weight_quantizers.38 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.34e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.gate_up_proj_weight_quantizers.39 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.13e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.gate_up_proj_weight_quantizers.40 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.29e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.gate_up_proj_weight_quantizers.41 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.gate_up_proj_weight_quantizers.42 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.50e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.gate_up_proj_weight_quantizers.43 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.gate_up_proj_weight_quantizers.44 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.gate_up_proj_weight_quantizers.45 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.59e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.gate_up_proj_weight_quantizers.46 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.gate_up_proj_weight_quantizers.47 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.gate_up_proj_weight_quantizers.48 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.46e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.gate_up_proj_weight_quantizers.49 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.07e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.gate_up_proj_weight_quantizers.50 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.gate_up_proj_weight_quantizers.51 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.16e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.gate_up_proj_weight_quantizers.52 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=5.55e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.gate_up_proj_weight_quantizers.53 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.24e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.gate_up_proj_weight_quantizers.54 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.23e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.gate_up_proj_weight_quantizers.55 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.15e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.gate_up_proj_weight_quantizers.56 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.gate_up_proj_weight_quantizers.57 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=4.47e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.gate_up_proj_weight_quantizers.58 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.21e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.gate_up_proj_weight_quantizers.59 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.gate_up_proj_weight_quantizers.60 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.gate_up_proj_weight_quantizers.61 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=5.94e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.gate_up_proj_weight_quantizers.62 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=7.42e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.gate_up_proj_weight_quantizers.63 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.gate_up_proj_weight_quantizers.64 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.gate_up_proj_weight_quantizers.65 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.13e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.gate_up_proj_weight_quantizers.66 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.13e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.gate_up_proj_weight_quantizers.67 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.22e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.gate_up_proj_weight_quantizers.68 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.42e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.gate_up_proj_weight_quantizers.69 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.gate_up_proj_weight_quantizers.70 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.14e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.gate_up_proj_weight_quantizers.71 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.gate_up_proj_weight_quantizers.72 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.gate_up_proj_weight_quantizers.73 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.20e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.gate_up_proj_weight_quantizers.74 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.06e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.gate_up_proj_weight_quantizers.75 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.gate_up_proj_weight_quantizers.76 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.19e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.gate_up_proj_weight_quantizers.77 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.gate_up_proj_weight_quantizers.78 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.24e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.gate_up_proj_weight_quantizers.79 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.gate_up_proj_weight_quantizers.80 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.67e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.gate_up_proj_weight_quantizers.81 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.gate_up_proj_weight_quantizers.82 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.gate_up_proj_weight_quantizers.83 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.13e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.gate_up_proj_weight_quantizers.84 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.gate_up_proj_weight_quantizers.85 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.42e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.gate_up_proj_weight_quantizers.86 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.gate_up_proj_weight_quantizers.87 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.36e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.gate_up_proj_weight_quantizers.88 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.gate_up_proj_weight_quantizers.89 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.48e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.gate_up_proj_weight_quantizers.90 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.40e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.gate_up_proj_weight_quantizers.91 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.31e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.gate_up_proj_weight_quantizers.92 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.16e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.gate_up_proj_weight_quantizers.93 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.gate_up_proj_weight_quantizers.94 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=7.03e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.gate_up_proj_weight_quantizers.95 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.24e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.gate_up_proj_weight_quantizers.96 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.15e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.gate_up_proj_weight_quantizers.97 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.gate_up_proj_weight_quantizers.98 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.07e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.gate_up_proj_weight_quantizers.99 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.gate_up_proj_weight_quantizers.100 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.20e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.gate_up_proj_weight_quantizers.101 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.gate_up_proj_weight_quantizers.102 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.29e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.gate_up_proj_weight_quantizers.103 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.64e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.gate_up_proj_weight_quantizers.104 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.46e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.gate_up_proj_weight_quantizers.105 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.67e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.gate_up_proj_weight_quantizers.106 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.gate_up_proj_weight_quantizers.107 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.gate_up_proj_weight_quantizers.108 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.gate_up_proj_weight_quantizers.109 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.gate_up_proj_weight_quantizers.110 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.gate_up_proj_weight_quantizers.111 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.gate_up_proj_weight_quantizers.112 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.76e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.gate_up_proj_weight_quantizers.113 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.gate_up_proj_weight_quantizers.114 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.85e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.gate_up_proj_weight_quantizers.115 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.65e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.gate_up_proj_weight_quantizers.116 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.gate_up_proj_weight_quantizers.117 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.36e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.gate_up_proj_weight_quantizers.118 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.35e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.gate_up_proj_weight_quantizers.119 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.22e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.gate_up_proj_weight_quantizers.120 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.19e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.gate_up_proj_weight_quantizers.121 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.31e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.gate_up_proj_weight_quantizers.122 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.52e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.gate_up_proj_weight_quantizers.123 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.32e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.gate_up_proj_weight_quantizers.124 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.gate_up_proj_weight_quantizers.125 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.gate_up_proj_weight_quantizers.126 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.gate_up_proj_weight_quantizers.127 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.gate_up_proj_weight_quantizers.128 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.gate_up_proj_weight_quantizers.129 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.49e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.gate_up_proj_weight_quantizers.130 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.78e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.gate_up_proj_weight_quantizers.131 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.gate_up_proj_weight_quantizers.132 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.16e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.gate_up_proj_weight_quantizers.133 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.15e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.gate_up_proj_weight_quantizers.134 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.87e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.gate_up_proj_weight_quantizers.135 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.gate_up_proj_weight_quantizers.136 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.gate_up_proj_weight_quantizers.137 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.17e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.gate_up_proj_weight_quantizers.138 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.gate_up_proj_weight_quantizers.139 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.gate_up_proj_weight_quantizers.140 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.05e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.gate_up_proj_weight_quantizers.141 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.gate_up_proj_weight_quantizers.142 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.gate_up_proj_weight_quantizers.143 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.31e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.gate_up_proj_weight_quantizers.144 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.gate_up_proj_weight_quantizers.145 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.22e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.gate_up_proj_weight_quantizers.146 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.gate_up_proj_weight_quantizers.147 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.54e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.gate_up_proj_weight_quantizers.148 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.52e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.gate_up_proj_weight_quantizers.149 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.gate_up_proj_weight_quantizers.150 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.42e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.gate_up_proj_weight_quantizers.151 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.gate_up_proj_weight_quantizers.152 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.gate_up_proj_weight_quantizers.153 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.gate_up_proj_weight_quantizers.154 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.23e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.gate_up_proj_weight_quantizers.155 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.89e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.gate_up_proj_weight_quantizers.156 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.gate_up_proj_weight_quantizers.157 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.16e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.gate_up_proj_weight_quantizers.158 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.gate_up_proj_weight_quantizers.159 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.gate_up_proj_weight_quantizers.160 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.13e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.gate_up_proj_weight_quantizers.161 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.gate_up_proj_weight_quantizers.162 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.18e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.gate_up_proj_weight_quantizers.163 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.gate_up_proj_weight_quantizers.164 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.gate_up_proj_weight_quantizers.165 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.22e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.gate_up_proj_weight_quantizers.166 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.43e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.gate_up_proj_weight_quantizers.167 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.gate_up_proj_weight_quantizers.168 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.47e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.gate_up_proj_weight_quantizers.169 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.gate_up_proj_weight_quantizers.170 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.46e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.gate_up_proj_weight_quantizers.171 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.28e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.gate_up_proj_weight_quantizers.172 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.gate_up_proj_weight_quantizers.173 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.14e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.gate_up_proj_weight_quantizers.174 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.gate_up_proj_weight_quantizers.175 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=6.37e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.gate_up_proj_weight_quantizers.176 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.gate_up_proj_weight_quantizers.177 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.21e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.gate_up_proj_weight_quantizers.178 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.21e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.gate_up_proj_weight_quantizers.179 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.64e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.gate_up_proj_weight_quantizers.180 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.gate_up_proj_weight_quantizers.181 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.gate_up_proj_weight_quantizers.182 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.gate_up_proj_weight_quantizers.183 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.gate_up_proj_weight_quantizers.184 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.24e+00 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.gate_up_proj_weight_quantizers.185 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.14e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.gate_up_proj_weight_quantizers.186 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.88e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.gate_up_proj_weight_quantizers.187 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.gate_up_proj_weight_quantizers.188 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.gate_up_proj_weight_quantizers.189 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.18e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.gate_up_proj_weight_quantizers.190 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.19e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.gate_up_proj_weight_quantizers.191 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.13e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.gate_up_proj_weight_quantizers.192 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.75e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.gate_up_proj_weight_quantizers.193 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.25e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.gate_up_proj_weight_quantizers.194 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.gate_up_proj_weight_quantizers.195 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.gate_up_proj_weight_quantizers.196 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.15e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.gate_up_proj_weight_quantizers.197 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=4.16e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.gate_up_proj_weight_quantizers.198 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.gate_up_proj_weight_quantizers.199 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.gate_up_proj_weight_quantizers.200 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.46e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.gate_up_proj_weight_quantizers.201 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.16e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.gate_up_proj_weight_quantizers.202 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.27e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.gate_up_proj_weight_quantizers.203 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.gate_up_proj_weight_quantizers.204 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.gate_up_proj_weight_quantizers.205 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.39e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.gate_up_proj_weight_quantizers.206 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.gate_up_proj_weight_quantizers.207 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.24e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.gate_up_proj_weight_quantizers.208 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.gate_up_proj_weight_quantizers.209 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.gate_up_proj_weight_quantizers.210 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.28e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.gate_up_proj_weight_quantizers.211 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.gate_up_proj_weight_quantizers.212 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.15e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.gate_up_proj_weight_quantizers.213 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.13e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.gate_up_proj_weight_quantizers.214 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.26e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.gate_up_proj_weight_quantizers.215 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.gate_up_proj_weight_quantizers.216 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.22e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.gate_up_proj_weight_quantizers.217 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.31e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.gate_up_proj_weight_quantizers.218 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.gate_up_proj_weight_quantizers.219 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.17e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.gate_up_proj_weight_quantizers.220 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.30e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.gate_up_proj_weight_quantizers.221 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.52e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.gate_up_proj_weight_quantizers.222 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.18e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.gate_up_proj_weight_quantizers.223 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.gate_up_proj_weight_quantizers.224 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.18e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.gate_up_proj_weight_quantizers.225 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.17e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.gate_up_proj_weight_quantizers.226 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.gate_up_proj_weight_quantizers.227 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.26e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.gate_up_proj_weight_quantizers.228 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.26e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.gate_up_proj_weight_quantizers.229 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.15e+00 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.gate_up_proj_weight_quantizers.230 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.gate_up_proj_weight_quantizers.231 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.68e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.gate_up_proj_weight_quantizers.232 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.26e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.gate_up_proj_weight_quantizers.233 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.56e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.gate_up_proj_weight_quantizers.234 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.22e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.gate_up_proj_weight_quantizers.235 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.gate_up_proj_weight_quantizers.236 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.gate_up_proj_weight_quantizers.237 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.70e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.gate_up_proj_weight_quantizers.238 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.gate_up_proj_weight_quantizers.239 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.gate_up_proj_weight_quantizers.240 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.gate_up_proj_weight_quantizers.241 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.23e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.gate_up_proj_weight_quantizers.242 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.14e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.gate_up_proj_weight_quantizers.243 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.gate_up_proj_weight_quantizers.244 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.gate_up_proj_weight_quantizers.245 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.gate_up_proj_weight_quantizers.246 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.45e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.gate_up_proj_weight_quantizers.247 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.21e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.gate_up_proj_weight_quantizers.248 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.24e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.gate_up_proj_weight_quantizers.249 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.gate_up_proj_weight_quantizers.250 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.25e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.gate_up_proj_weight_quantizers.251 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.gate_up_proj_weight_quantizers.252 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.17e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.gate_up_proj_weight_quantizers.253 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.gate_up_proj_weight_quantizers.254 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.gate_up_proj_weight_quantizers.255 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=4.02e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.down_proj_input_quantizer TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.80e+01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.down_proj_weight_quantizers.0 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.down_proj_weight_quantizers.1 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.54e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.down_proj_weight_quantizers.2 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.29e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.down_proj_weight_quantizers.3 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.down_proj_weight_quantizers.4 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.down_proj_weight_quantizers.5 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.down_proj_weight_quantizers.6 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.41e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.down_proj_weight_quantizers.7 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.43e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.down_proj_weight_quantizers.8 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.down_proj_weight_quantizers.9 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.down_proj_weight_quantizers.10 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.down_proj_weight_quantizers.11 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.down_proj_weight_quantizers.12 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.down_proj_weight_quantizers.13 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.29e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.down_proj_weight_quantizers.14 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.down_proj_weight_quantizers.15 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.down_proj_weight_quantizers.16 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.down_proj_weight_quantizers.17 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.down_proj_weight_quantizers.18 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.down_proj_weight_quantizers.19 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.down_proj_weight_quantizers.20 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.55e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.down_proj_weight_quantizers.21 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.down_proj_weight_quantizers.22 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.down_proj_weight_quantizers.23 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.down_proj_weight_quantizers.24 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.down_proj_weight_quantizers.25 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.down_proj_weight_quantizers.26 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.down_proj_weight_quantizers.27 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.down_proj_weight_quantizers.28 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.down_proj_weight_quantizers.29 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.down_proj_weight_quantizers.30 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.down_proj_weight_quantizers.31 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.down_proj_weight_quantizers.32 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.15e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.down_proj_weight_quantizers.33 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.down_proj_weight_quantizers.34 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.16e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.down_proj_weight_quantizers.35 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.down_proj_weight_quantizers.36 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.down_proj_weight_quantizers.37 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.25e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.down_proj_weight_quantizers.38 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.down_proj_weight_quantizers.39 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.down_proj_weight_quantizers.40 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.84e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.down_proj_weight_quantizers.41 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.down_proj_weight_quantizers.42 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.15e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.down_proj_weight_quantizers.43 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.down_proj_weight_quantizers.44 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.down_proj_weight_quantizers.45 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.20e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.down_proj_weight_quantizers.46 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.down_proj_weight_quantizers.47 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.down_proj_weight_quantizers.48 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.down_proj_weight_quantizers.49 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.32e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.down_proj_weight_quantizers.50 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.14e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.down_proj_weight_quantizers.51 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.down_proj_weight_quantizers.52 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.82e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.down_proj_weight_quantizers.53 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.down_proj_weight_quantizers.54 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.down_proj_weight_quantizers.55 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.down_proj_weight_quantizers.56 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.down_proj_weight_quantizers.57 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.down_proj_weight_quantizers.58 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.39e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.down_proj_weight_quantizers.59 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.down_proj_weight_quantizers.60 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.down_proj_weight_quantizers.61 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.40e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.down_proj_weight_quantizers.62 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.29e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.down_proj_weight_quantizers.63 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.down_proj_weight_quantizers.64 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.down_proj_weight_quantizers.65 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.down_proj_weight_quantizers.66 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.36e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.down_proj_weight_quantizers.67 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.14e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.down_proj_weight_quantizers.68 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.down_proj_weight_quantizers.69 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.down_proj_weight_quantizers.70 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.down_proj_weight_quantizers.71 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.down_proj_weight_quantizers.72 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.down_proj_weight_quantizers.73 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.down_proj_weight_quantizers.74 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.down_proj_weight_quantizers.75 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.down_proj_weight_quantizers.76 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.down_proj_weight_quantizers.77 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.19e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.down_proj_weight_quantizers.78 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.27e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.down_proj_weight_quantizers.79 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.down_proj_weight_quantizers.80 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.down_proj_weight_quantizers.81 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.down_proj_weight_quantizers.82 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.56e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.down_proj_weight_quantizers.83 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.14e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.down_proj_weight_quantizers.84 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.down_proj_weight_quantizers.85 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.15e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.down_proj_weight_quantizers.86 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.down_proj_weight_quantizers.87 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.down_proj_weight_quantizers.88 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.down_proj_weight_quantizers.89 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.down_proj_weight_quantizers.90 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.27e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.down_proj_weight_quantizers.91 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.down_proj_weight_quantizers.92 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.down_proj_weight_quantizers.93 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.down_proj_weight_quantizers.94 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.90e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.down_proj_weight_quantizers.95 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.down_proj_weight_quantizers.96 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.down_proj_weight_quantizers.97 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.17e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.down_proj_weight_quantizers.98 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.00e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.down_proj_weight_quantizers.99 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.down_proj_weight_quantizers.100 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.17e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.down_proj_weight_quantizers.101 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.down_proj_weight_quantizers.102 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.down_proj_weight_quantizers.103 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.down_proj_weight_quantizers.104 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.down_proj_weight_quantizers.105 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.97e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.down_proj_weight_quantizers.106 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.down_proj_weight_quantizers.107 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.down_proj_weight_quantizers.108 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.down_proj_weight_quantizers.109 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.down_proj_weight_quantizers.110 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.down_proj_weight_quantizers.111 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.18e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.down_proj_weight_quantizers.112 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.down_proj_weight_quantizers.113 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.down_proj_weight_quantizers.114 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.down_proj_weight_quantizers.115 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.down_proj_weight_quantizers.116 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.down_proj_weight_quantizers.117 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.37e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.down_proj_weight_quantizers.118 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.down_proj_weight_quantizers.119 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.down_proj_weight_quantizers.120 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.18e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.down_proj_weight_quantizers.121 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.73e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.down_proj_weight_quantizers.122 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.down_proj_weight_quantizers.123 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.22e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.down_proj_weight_quantizers.124 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.down_proj_weight_quantizers.125 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.down_proj_weight_quantizers.126 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.down_proj_weight_quantizers.127 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.down_proj_weight_quantizers.128 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.down_proj_weight_quantizers.129 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.down_proj_weight_quantizers.130 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.down_proj_weight_quantizers.131 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.down_proj_weight_quantizers.132 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.down_proj_weight_quantizers.133 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.down_proj_weight_quantizers.134 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.18e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.down_proj_weight_quantizers.135 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.15e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.down_proj_weight_quantizers.136 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.down_proj_weight_quantizers.137 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.down_proj_weight_quantizers.138 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.down_proj_weight_quantizers.139 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.down_proj_weight_quantizers.140 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.down_proj_weight_quantizers.141 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.down_proj_weight_quantizers.142 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.down_proj_weight_quantizers.143 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.down_proj_weight_quantizers.144 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.down_proj_weight_quantizers.145 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.down_proj_weight_quantizers.146 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.down_proj_weight_quantizers.147 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.46e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.down_proj_weight_quantizers.148 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.down_proj_weight_quantizers.149 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.46e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.down_proj_weight_quantizers.150 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.down_proj_weight_quantizers.151 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.down_proj_weight_quantizers.152 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.22e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.down_proj_weight_quantizers.153 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.down_proj_weight_quantizers.154 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.down_proj_weight_quantizers.155 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.down_proj_weight_quantizers.156 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.down_proj_weight_quantizers.157 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.13e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.down_proj_weight_quantizers.158 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.down_proj_weight_quantizers.159 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.down_proj_weight_quantizers.160 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.down_proj_weight_quantizers.161 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.down_proj_weight_quantizers.162 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.down_proj_weight_quantizers.163 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.down_proj_weight_quantizers.164 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.down_proj_weight_quantizers.165 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.down_proj_weight_quantizers.166 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.down_proj_weight_quantizers.167 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.down_proj_weight_quantizers.168 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.down_proj_weight_quantizers.169 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.down_proj_weight_quantizers.170 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.down_proj_weight_quantizers.171 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.down_proj_weight_quantizers.172 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.44e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.down_proj_weight_quantizers.173 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.down_proj_weight_quantizers.174 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.down_proj_weight_quantizers.175 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.44e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.down_proj_weight_quantizers.176 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.down_proj_weight_quantizers.177 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.33e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.down_proj_weight_quantizers.178 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.down_proj_weight_quantizers.179 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.down_proj_weight_quantizers.180 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.down_proj_weight_quantizers.181 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.down_proj_weight_quantizers.182 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.13e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.down_proj_weight_quantizers.183 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.42e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.down_proj_weight_quantizers.184 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.85e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.down_proj_weight_quantizers.185 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.down_proj_weight_quantizers.186 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.14e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.down_proj_weight_quantizers.187 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.down_proj_weight_quantizers.188 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.down_proj_weight_quantizers.189 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.down_proj_weight_quantizers.190 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.down_proj_weight_quantizers.191 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.down_proj_weight_quantizers.192 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.84e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.down_proj_weight_quantizers.193 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.down_proj_weight_quantizers.194 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.down_proj_weight_quantizers.195 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.down_proj_weight_quantizers.196 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.down_proj_weight_quantizers.197 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.87e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.down_proj_weight_quantizers.198 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.down_proj_weight_quantizers.199 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.down_proj_weight_quantizers.200 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.down_proj_weight_quantizers.201 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.down_proj_weight_quantizers.202 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.down_proj_weight_quantizers.203 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.down_proj_weight_quantizers.204 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.down_proj_weight_quantizers.205 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.68e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.down_proj_weight_quantizers.206 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.down_proj_weight_quantizers.207 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.down_proj_weight_quantizers.208 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.down_proj_weight_quantizers.209 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.56e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.down_proj_weight_quantizers.210 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.down_proj_weight_quantizers.211 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.down_proj_weight_quantizers.212 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.down_proj_weight_quantizers.213 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.down_proj_weight_quantizers.214 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.down_proj_weight_quantizers.215 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.down_proj_weight_quantizers.216 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.down_proj_weight_quantizers.217 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.47e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.down_proj_weight_quantizers.218 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.down_proj_weight_quantizers.219 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.20e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.down_proj_weight_quantizers.220 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.55e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.down_proj_weight_quantizers.221 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.72e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.down_proj_weight_quantizers.222 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.down_proj_weight_quantizers.223 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.down_proj_weight_quantizers.224 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.down_proj_weight_quantizers.225 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.down_proj_weight_quantizers.226 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.down_proj_weight_quantizers.227 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.down_proj_weight_quantizers.228 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.down_proj_weight_quantizers.229 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.25e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.down_proj_weight_quantizers.230 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.down_proj_weight_quantizers.231 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.down_proj_weight_quantizers.232 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.down_proj_weight_quantizers.233 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.down_proj_weight_quantizers.234 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.down_proj_weight_quantizers.235 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.down_proj_weight_quantizers.236 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.down_proj_weight_quantizers.237 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.39e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.down_proj_weight_quantizers.238 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.down_proj_weight_quantizers.239 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.41e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.down_proj_weight_quantizers.240 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.down_proj_weight_quantizers.241 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.down_proj_weight_quantizers.242 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.down_proj_weight_quantizers.243 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.down_proj_weight_quantizers.244 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.down_proj_weight_quantizers.245 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.down_proj_weight_quantizers.246 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.34e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.down_proj_weight_quantizers.247 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.31e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.down_proj_weight_quantizers.248 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.26e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.down_proj_weight_quantizers.249 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.down_proj_weight_quantizers.250 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.down_proj_weight_quantizers.251 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.down_proj_weight_quantizers.252 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.down_proj_weight_quantizers.253 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.down_proj_weight_quantizers.254 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.down_proj_weight_quantizers.255 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.80e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.shared_experts.gate_proj.input_quantizer TensorQuantizer(disabled) +model.layers.70.mlp.shared_experts.gate_proj.output_quantizer TensorQuantizer(disabled) +model.layers.70.mlp.shared_experts.gate_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.70.mlp.shared_experts.up_proj.input_quantizer TensorQuantizer(disabled) +model.layers.70.mlp.shared_experts.up_proj.output_quantizer TensorQuantizer(disabled) +model.layers.70.mlp.shared_experts.up_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.70.mlp.shared_experts.down_proj.input_quantizer TensorQuantizer(disabled) +model.layers.70.mlp.shared_experts.down_proj.output_quantizer TensorQuantizer(disabled) +model.layers.70.mlp.shared_experts.down_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.71.self_attn.q_a_proj.input_quantizer TensorQuantizer(disabled) +model.layers.71.self_attn.q_a_proj.output_quantizer TensorQuantizer(disabled) +model.layers.71.self_attn.q_a_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.71.self_attn.q_b_proj.input_quantizer TensorQuantizer(disabled) +model.layers.71.self_attn.q_b_proj.output_quantizer TensorQuantizer(disabled) +model.layers.71.self_attn.q_b_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.71.self_attn.kv_a_proj_with_mqa.input_quantizer TensorQuantizer(disabled) +model.layers.71.self_attn.kv_a_proj_with_mqa.output_quantizer TensorQuantizer(disabled) +model.layers.71.self_attn.kv_a_proj_with_mqa.weight_quantizer TensorQuantizer(disabled) +model.layers.71.self_attn.kv_b_proj.input_quantizer TensorQuantizer(disabled) +model.layers.71.self_attn.kv_b_proj.output_quantizer TensorQuantizer(disabled) +model.layers.71.self_attn.kv_b_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.71.self_attn.o_proj.input_quantizer TensorQuantizer(disabled) +model.layers.71.self_attn.o_proj.output_quantizer TensorQuantizer(disabled) +model.layers.71.self_attn.o_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.71.self_attn.q_bmm_quantizer TensorQuantizer(disabled) +model.layers.71.self_attn.k_bmm_quantizer TensorQuantizer((4, 3) bit fake per-tensor amax=4.48e+02(const) calibrator=MaxCalibrator quant) +model.layers.71.self_attn.v_bmm_quantizer TensorQuantizer((4, 3) bit fake per-tensor amax=4.48e+02(const) calibrator=MaxCalibrator quant) +model.layers.71.self_attn.p_bmm_quantizer TensorQuantizer(disabled) +model.layers.71.mlp.experts.gate_up_proj_input_quantizer TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.36e+00 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.gate_up_proj_weight_quantizers.0 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.gate_up_proj_weight_quantizers.1 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.28e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.gate_up_proj_weight_quantizers.2 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.17e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.gate_up_proj_weight_quantizers.3 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.12e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.gate_up_proj_weight_quantizers.4 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.19e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.gate_up_proj_weight_quantizers.5 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.35e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.gate_up_proj_weight_quantizers.6 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.19e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.gate_up_proj_weight_quantizers.7 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.46e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.gate_up_proj_weight_quantizers.8 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.gate_up_proj_weight_quantizers.9 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.28e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.gate_up_proj_weight_quantizers.10 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.gate_up_proj_weight_quantizers.11 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.97e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.gate_up_proj_weight_quantizers.12 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.21e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.gate_up_proj_weight_quantizers.13 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.gate_up_proj_weight_quantizers.14 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.89e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.gate_up_proj_weight_quantizers.15 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.gate_up_proj_weight_quantizers.16 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.24e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.gate_up_proj_weight_quantizers.17 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.gate_up_proj_weight_quantizers.18 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.gate_up_proj_weight_quantizers.19 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.gate_up_proj_weight_quantizers.20 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=6.21e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.gate_up_proj_weight_quantizers.21 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.19e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.gate_up_proj_weight_quantizers.22 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.27e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.gate_up_proj_weight_quantizers.23 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.79e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.gate_up_proj_weight_quantizers.24 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.56e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.gate_up_proj_weight_quantizers.25 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.17e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.gate_up_proj_weight_quantizers.26 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.25e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.gate_up_proj_weight_quantizers.27 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.91e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.gate_up_proj_weight_quantizers.28 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.gate_up_proj_weight_quantizers.29 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.gate_up_proj_weight_quantizers.30 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.15e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.gate_up_proj_weight_quantizers.31 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.gate_up_proj_weight_quantizers.32 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=5.27e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.gate_up_proj_weight_quantizers.33 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.32e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.gate_up_proj_weight_quantizers.34 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.gate_up_proj_weight_quantizers.35 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.32e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.gate_up_proj_weight_quantizers.36 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.23e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.gate_up_proj_weight_quantizers.37 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=4.84e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.gate_up_proj_weight_quantizers.38 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.37e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.gate_up_proj_weight_quantizers.39 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.gate_up_proj_weight_quantizers.40 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.gate_up_proj_weight_quantizers.41 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.gate_up_proj_weight_quantizers.42 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.23e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.gate_up_proj_weight_quantizers.43 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.gate_up_proj_weight_quantizers.44 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.23e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.gate_up_proj_weight_quantizers.45 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.gate_up_proj_weight_quantizers.46 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.15e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.gate_up_proj_weight_quantizers.47 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.gate_up_proj_weight_quantizers.48 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=4.28e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.gate_up_proj_weight_quantizers.49 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.gate_up_proj_weight_quantizers.50 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.gate_up_proj_weight_quantizers.51 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.22e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.gate_up_proj_weight_quantizers.52 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.gate_up_proj_weight_quantizers.53 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.gate_up_proj_weight_quantizers.54 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.16e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.gate_up_proj_weight_quantizers.55 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.gate_up_proj_weight_quantizers.56 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.gate_up_proj_weight_quantizers.57 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.13e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.gate_up_proj_weight_quantizers.58 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.15e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.gate_up_proj_weight_quantizers.59 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.18e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.gate_up_proj_weight_quantizers.60 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.35e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.gate_up_proj_weight_quantizers.61 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.43e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.gate_up_proj_weight_quantizers.62 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.33e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.gate_up_proj_weight_quantizers.63 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.gate_up_proj_weight_quantizers.64 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.60e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.gate_up_proj_weight_quantizers.65 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.17e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.gate_up_proj_weight_quantizers.66 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.41e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.gate_up_proj_weight_quantizers.67 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.gate_up_proj_weight_quantizers.68 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.25e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.gate_up_proj_weight_quantizers.69 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.44e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.gate_up_proj_weight_quantizers.70 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.18e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.gate_up_proj_weight_quantizers.71 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.gate_up_proj_weight_quantizers.72 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.gate_up_proj_weight_quantizers.73 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.gate_up_proj_weight_quantizers.74 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.14e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.gate_up_proj_weight_quantizers.75 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.44e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.gate_up_proj_weight_quantizers.76 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.33e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.gate_up_proj_weight_quantizers.77 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.13e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.gate_up_proj_weight_quantizers.78 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.77e+00 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.gate_up_proj_weight_quantizers.79 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.81e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.gate_up_proj_weight_quantizers.80 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.22e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.gate_up_proj_weight_quantizers.81 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.gate_up_proj_weight_quantizers.82 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.23e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.gate_up_proj_weight_quantizers.83 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.gate_up_proj_weight_quantizers.84 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.gate_up_proj_weight_quantizers.85 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.48e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.gate_up_proj_weight_quantizers.86 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.83e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.gate_up_proj_weight_quantizers.87 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.24e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.gate_up_proj_weight_quantizers.88 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.26e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.gate_up_proj_weight_quantizers.89 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.gate_up_proj_weight_quantizers.90 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.gate_up_proj_weight_quantizers.91 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=5.04e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.gate_up_proj_weight_quantizers.92 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.13e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.gate_up_proj_weight_quantizers.93 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.13e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.gate_up_proj_weight_quantizers.94 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.32e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.gate_up_proj_weight_quantizers.95 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=7.46e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.gate_up_proj_weight_quantizers.96 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.gate_up_proj_weight_quantizers.97 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.20e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.gate_up_proj_weight_quantizers.98 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.39e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.gate_up_proj_weight_quantizers.99 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.gate_up_proj_weight_quantizers.100 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.23e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.gate_up_proj_weight_quantizers.101 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.gate_up_proj_weight_quantizers.102 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.66e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.gate_up_proj_weight_quantizers.103 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.26e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.gate_up_proj_weight_quantizers.104 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.32e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.gate_up_proj_weight_quantizers.105 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.gate_up_proj_weight_quantizers.106 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.30e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.gate_up_proj_weight_quantizers.107 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.gate_up_proj_weight_quantizers.108 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.gate_up_proj_weight_quantizers.109 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.18e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.gate_up_proj_weight_quantizers.110 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=6.56e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.gate_up_proj_weight_quantizers.111 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.47e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.gate_up_proj_weight_quantizers.112 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.43e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.gate_up_proj_weight_quantizers.113 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.gate_up_proj_weight_quantizers.114 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.42e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.gate_up_proj_weight_quantizers.115 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.gate_up_proj_weight_quantizers.116 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.17e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.gate_up_proj_weight_quantizers.117 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=4.34e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.gate_up_proj_weight_quantizers.118 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.gate_up_proj_weight_quantizers.119 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.30e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.gate_up_proj_weight_quantizers.120 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.49e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.gate_up_proj_weight_quantizers.121 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.gate_up_proj_weight_quantizers.122 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.25e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.gate_up_proj_weight_quantizers.123 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.15e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.gate_up_proj_weight_quantizers.124 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.11e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.gate_up_proj_weight_quantizers.125 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.49e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.gate_up_proj_weight_quantizers.126 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.20e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.gate_up_proj_weight_quantizers.127 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.gate_up_proj_weight_quantizers.128 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.gate_up_proj_weight_quantizers.129 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.36e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.gate_up_proj_weight_quantizers.130 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.21e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.gate_up_proj_weight_quantizers.131 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.gate_up_proj_weight_quantizers.132 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.22e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.gate_up_proj_weight_quantizers.133 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.gate_up_proj_weight_quantizers.134 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.14e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.gate_up_proj_weight_quantizers.135 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.25e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.gate_up_proj_weight_quantizers.136 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.gate_up_proj_weight_quantizers.137 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.gate_up_proj_weight_quantizers.138 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.gate_up_proj_weight_quantizers.139 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.gate_up_proj_weight_quantizers.140 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.20e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.gate_up_proj_weight_quantizers.141 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.34e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.gate_up_proj_weight_quantizers.142 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.58e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.gate_up_proj_weight_quantizers.143 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.gate_up_proj_weight_quantizers.144 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.57e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.gate_up_proj_weight_quantizers.145 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.gate_up_proj_weight_quantizers.146 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.23e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.gate_up_proj_weight_quantizers.147 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.15e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.gate_up_proj_weight_quantizers.148 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.05e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.gate_up_proj_weight_quantizers.149 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.13e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.gate_up_proj_weight_quantizers.150 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.gate_up_proj_weight_quantizers.151 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.gate_up_proj_weight_quantizers.152 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.60e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.gate_up_proj_weight_quantizers.153 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.gate_up_proj_weight_quantizers.154 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.36e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.gate_up_proj_weight_quantizers.155 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.gate_up_proj_weight_quantizers.156 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.60e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.gate_up_proj_weight_quantizers.157 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.45e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.gate_up_proj_weight_quantizers.158 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.gate_up_proj_weight_quantizers.159 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.43e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.gate_up_proj_weight_quantizers.160 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.gate_up_proj_weight_quantizers.161 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.93e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.gate_up_proj_weight_quantizers.162 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.19e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.gate_up_proj_weight_quantizers.163 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.18e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.gate_up_proj_weight_quantizers.164 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.25e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.gate_up_proj_weight_quantizers.165 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.33e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.gate_up_proj_weight_quantizers.166 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.gate_up_proj_weight_quantizers.167 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.46e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.gate_up_proj_weight_quantizers.168 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.49e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.gate_up_proj_weight_quantizers.169 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.gate_up_proj_weight_quantizers.170 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.41e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.gate_up_proj_weight_quantizers.171 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.gate_up_proj_weight_quantizers.172 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.73e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.gate_up_proj_weight_quantizers.173 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.16e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.gate_up_proj_weight_quantizers.174 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.24e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.gate_up_proj_weight_quantizers.175 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.18e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.gate_up_proj_weight_quantizers.176 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.23e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.gate_up_proj_weight_quantizers.177 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.gate_up_proj_weight_quantizers.178 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.41e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.gate_up_proj_weight_quantizers.179 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.gate_up_proj_weight_quantizers.180 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.gate_up_proj_weight_quantizers.181 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.gate_up_proj_weight_quantizers.182 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.36e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.gate_up_proj_weight_quantizers.183 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.gate_up_proj_weight_quantizers.184 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.gate_up_proj_weight_quantizers.185 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.91e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.gate_up_proj_weight_quantizers.186 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.13e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.gate_up_proj_weight_quantizers.187 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.gate_up_proj_weight_quantizers.188 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.gate_up_proj_weight_quantizers.189 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.18e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.gate_up_proj_weight_quantizers.190 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.13e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.gate_up_proj_weight_quantizers.191 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=4.26e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.gate_up_proj_weight_quantizers.192 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.34e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.gate_up_proj_weight_quantizers.193 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.gate_up_proj_weight_quantizers.194 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.35e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.gate_up_proj_weight_quantizers.195 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.gate_up_proj_weight_quantizers.196 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.42e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.gate_up_proj_weight_quantizers.197 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.47e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.gate_up_proj_weight_quantizers.198 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.24e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.gate_up_proj_weight_quantizers.199 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.gate_up_proj_weight_quantizers.200 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.23e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.gate_up_proj_weight_quantizers.201 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=7.30e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.gate_up_proj_weight_quantizers.202 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.25e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.gate_up_proj_weight_quantizers.203 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.gate_up_proj_weight_quantizers.204 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.gate_up_proj_weight_quantizers.205 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.32e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.gate_up_proj_weight_quantizers.206 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.16e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.gate_up_proj_weight_quantizers.207 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.14e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.gate_up_proj_weight_quantizers.208 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.33e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.gate_up_proj_weight_quantizers.209 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.35e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.gate_up_proj_weight_quantizers.210 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.90e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.gate_up_proj_weight_quantizers.211 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.30e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.gate_up_proj_weight_quantizers.212 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.28e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.gate_up_proj_weight_quantizers.213 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.gate_up_proj_weight_quantizers.214 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=4.18e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.gate_up_proj_weight_quantizers.215 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.38e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.gate_up_proj_weight_quantizers.216 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.34e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.gate_up_proj_weight_quantizers.217 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.gate_up_proj_weight_quantizers.218 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.35e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.gate_up_proj_weight_quantizers.219 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.gate_up_proj_weight_quantizers.220 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.gate_up_proj_weight_quantizers.221 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.18e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.gate_up_proj_weight_quantizers.222 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.15e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.gate_up_proj_weight_quantizers.223 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.80e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.gate_up_proj_weight_quantizers.224 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.gate_up_proj_weight_quantizers.225 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.73e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.gate_up_proj_weight_quantizers.226 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.13e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.gate_up_proj_weight_quantizers.227 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.15e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.gate_up_proj_weight_quantizers.228 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.29e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.gate_up_proj_weight_quantizers.229 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.gate_up_proj_weight_quantizers.230 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.gate_up_proj_weight_quantizers.231 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.34e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.gate_up_proj_weight_quantizers.232 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.gate_up_proj_weight_quantizers.233 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.gate_up_proj_weight_quantizers.234 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.gate_up_proj_weight_quantizers.235 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.gate_up_proj_weight_quantizers.236 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.gate_up_proj_weight_quantizers.237 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.gate_up_proj_weight_quantizers.238 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.18e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.gate_up_proj_weight_quantizers.239 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.23e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.gate_up_proj_weight_quantizers.240 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.13e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.gate_up_proj_weight_quantizers.241 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.32e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.gate_up_proj_weight_quantizers.242 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.45e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.gate_up_proj_weight_quantizers.243 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.14e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.gate_up_proj_weight_quantizers.244 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.35e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.gate_up_proj_weight_quantizers.245 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.67e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.gate_up_proj_weight_quantizers.246 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.21e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.gate_up_proj_weight_quantizers.247 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=7.46e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.gate_up_proj_weight_quantizers.248 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.gate_up_proj_weight_quantizers.249 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.13e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.gate_up_proj_weight_quantizers.250 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.13e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.gate_up_proj_weight_quantizers.251 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.21e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.gate_up_proj_weight_quantizers.252 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.46e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.gate_up_proj_weight_quantizers.253 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.32e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.gate_up_proj_weight_quantizers.254 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.gate_up_proj_weight_quantizers.255 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.23e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.down_proj_input_quantizer TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.95e+01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.down_proj_weight_quantizers.0 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.down_proj_weight_quantizers.1 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.down_proj_weight_quantizers.2 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.down_proj_weight_quantizers.3 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.down_proj_weight_quantizers.4 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.down_proj_weight_quantizers.5 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.down_proj_weight_quantizers.6 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.down_proj_weight_quantizers.7 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.down_proj_weight_quantizers.8 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.down_proj_weight_quantizers.9 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.down_proj_weight_quantizers.10 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.down_proj_weight_quantizers.11 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=4.14e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.down_proj_weight_quantizers.12 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.down_proj_weight_quantizers.13 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.65e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.down_proj_weight_quantizers.14 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.down_proj_weight_quantizers.15 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.down_proj_weight_quantizers.16 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.down_proj_weight_quantizers.17 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.down_proj_weight_quantizers.18 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.down_proj_weight_quantizers.19 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.down_proj_weight_quantizers.20 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.79e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.down_proj_weight_quantizers.21 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.19e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.down_proj_weight_quantizers.22 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.down_proj_weight_quantizers.23 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.30e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.down_proj_weight_quantizers.24 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.down_proj_weight_quantizers.25 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.down_proj_weight_quantizers.26 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.down_proj_weight_quantizers.27 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.95e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.down_proj_weight_quantizers.28 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.down_proj_weight_quantizers.29 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.down_proj_weight_quantizers.30 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.down_proj_weight_quantizers.31 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.down_proj_weight_quantizers.32 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.41e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.down_proj_weight_quantizers.33 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.down_proj_weight_quantizers.34 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.59e-02 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.down_proj_weight_quantizers.35 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.down_proj_weight_quantizers.36 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.down_proj_weight_quantizers.37 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.60e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.down_proj_weight_quantizers.38 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.down_proj_weight_quantizers.39 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.down_proj_weight_quantizers.40 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.down_proj_weight_quantizers.41 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.down_proj_weight_quantizers.42 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.down_proj_weight_quantizers.43 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.down_proj_weight_quantizers.44 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.down_proj_weight_quantizers.45 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.down_proj_weight_quantizers.46 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.58e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.down_proj_weight_quantizers.47 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.down_proj_weight_quantizers.48 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.41e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.down_proj_weight_quantizers.49 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.down_proj_weight_quantizers.50 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.down_proj_weight_quantizers.51 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.down_proj_weight_quantizers.52 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.down_proj_weight_quantizers.53 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.13e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.down_proj_weight_quantizers.54 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.16e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.down_proj_weight_quantizers.55 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.down_proj_weight_quantizers.56 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.46e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.down_proj_weight_quantizers.57 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.down_proj_weight_quantizers.58 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.down_proj_weight_quantizers.59 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.down_proj_weight_quantizers.60 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.down_proj_weight_quantizers.61 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.down_proj_weight_quantizers.62 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.down_proj_weight_quantizers.63 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.down_proj_weight_quantizers.64 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.down_proj_weight_quantizers.65 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.down_proj_weight_quantizers.66 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.down_proj_weight_quantizers.67 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.down_proj_weight_quantizers.68 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.down_proj_weight_quantizers.69 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.down_proj_weight_quantizers.70 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.75e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.down_proj_weight_quantizers.71 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.down_proj_weight_quantizers.72 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.down_proj_weight_quantizers.73 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.down_proj_weight_quantizers.74 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.down_proj_weight_quantizers.75 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.down_proj_weight_quantizers.76 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.13e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.down_proj_weight_quantizers.77 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.down_proj_weight_quantizers.78 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.46e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.down_proj_weight_quantizers.79 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.down_proj_weight_quantizers.80 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.down_proj_weight_quantizers.81 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.down_proj_weight_quantizers.82 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.down_proj_weight_quantizers.83 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.down_proj_weight_quantizers.84 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.down_proj_weight_quantizers.85 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.down_proj_weight_quantizers.86 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.43e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.down_proj_weight_quantizers.87 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.down_proj_weight_quantizers.88 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.down_proj_weight_quantizers.89 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.down_proj_weight_quantizers.90 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.down_proj_weight_quantizers.91 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.64e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.down_proj_weight_quantizers.92 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.57e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.down_proj_weight_quantizers.93 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.down_proj_weight_quantizers.94 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.down_proj_weight_quantizers.95 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.44e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.down_proj_weight_quantizers.96 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.down_proj_weight_quantizers.97 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.down_proj_weight_quantizers.98 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.down_proj_weight_quantizers.99 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.down_proj_weight_quantizers.100 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.down_proj_weight_quantizers.101 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.down_proj_weight_quantizers.102 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.44e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.down_proj_weight_quantizers.103 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.down_proj_weight_quantizers.104 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.52e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.down_proj_weight_quantizers.105 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.down_proj_weight_quantizers.106 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.down_proj_weight_quantizers.107 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.down_proj_weight_quantizers.108 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.14e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.down_proj_weight_quantizers.109 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.down_proj_weight_quantizers.110 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.85e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.down_proj_weight_quantizers.111 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.down_proj_weight_quantizers.112 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.down_proj_weight_quantizers.113 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.down_proj_weight_quantizers.114 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.25e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.down_proj_weight_quantizers.115 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.down_proj_weight_quantizers.116 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.down_proj_weight_quantizers.117 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.93e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.down_proj_weight_quantizers.118 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.down_proj_weight_quantizers.119 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.down_proj_weight_quantizers.120 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.down_proj_weight_quantizers.121 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.down_proj_weight_quantizers.122 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.down_proj_weight_quantizers.123 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.72e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.down_proj_weight_quantizers.124 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.37e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.down_proj_weight_quantizers.125 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.49e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.down_proj_weight_quantizers.126 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.down_proj_weight_quantizers.127 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.down_proj_weight_quantizers.128 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.down_proj_weight_quantizers.129 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.down_proj_weight_quantizers.130 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.down_proj_weight_quantizers.131 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.down_proj_weight_quantizers.132 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.33e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.down_proj_weight_quantizers.133 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.down_proj_weight_quantizers.134 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.down_proj_weight_quantizers.135 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.down_proj_weight_quantizers.136 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.01e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.down_proj_weight_quantizers.137 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.down_proj_weight_quantizers.138 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.down_proj_weight_quantizers.139 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.down_proj_weight_quantizers.140 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.down_proj_weight_quantizers.141 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.23e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.down_proj_weight_quantizers.142 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.down_proj_weight_quantizers.143 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.down_proj_weight_quantizers.144 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.down_proj_weight_quantizers.145 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.down_proj_weight_quantizers.146 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.24e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.down_proj_weight_quantizers.147 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.down_proj_weight_quantizers.148 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.33e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.down_proj_weight_quantizers.149 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.down_proj_weight_quantizers.150 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.down_proj_weight_quantizers.151 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.down_proj_weight_quantizers.152 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.down_proj_weight_quantizers.153 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.down_proj_weight_quantizers.154 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.down_proj_weight_quantizers.155 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.down_proj_weight_quantizers.156 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.down_proj_weight_quantizers.157 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.40e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.down_proj_weight_quantizers.158 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.down_proj_weight_quantizers.159 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.down_proj_weight_quantizers.160 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.down_proj_weight_quantizers.161 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.down_proj_weight_quantizers.162 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.down_proj_weight_quantizers.163 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.38e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.down_proj_weight_quantizers.164 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.16e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.down_proj_weight_quantizers.165 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.down_proj_weight_quantizers.166 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.down_proj_weight_quantizers.167 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.down_proj_weight_quantizers.168 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.down_proj_weight_quantizers.169 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.down_proj_weight_quantizers.170 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.down_proj_weight_quantizers.171 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.down_proj_weight_quantizers.172 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.14e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.down_proj_weight_quantizers.173 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.down_proj_weight_quantizers.174 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.down_proj_weight_quantizers.175 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.down_proj_weight_quantizers.176 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.down_proj_weight_quantizers.177 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.down_proj_weight_quantizers.178 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.72e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.down_proj_weight_quantizers.179 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.down_proj_weight_quantizers.180 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.down_proj_weight_quantizers.181 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.down_proj_weight_quantizers.182 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.down_proj_weight_quantizers.183 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.down_proj_weight_quantizers.184 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.10e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.down_proj_weight_quantizers.185 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.70e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.down_proj_weight_quantizers.186 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.down_proj_weight_quantizers.187 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.down_proj_weight_quantizers.188 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.down_proj_weight_quantizers.189 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.down_proj_weight_quantizers.190 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.down_proj_weight_quantizers.191 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.24e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.down_proj_weight_quantizers.192 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.20e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.down_proj_weight_quantizers.193 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.down_proj_weight_quantizers.194 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.down_proj_weight_quantizers.195 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.down_proj_weight_quantizers.196 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.47e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.down_proj_weight_quantizers.197 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.down_proj_weight_quantizers.198 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.down_proj_weight_quantizers.199 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.down_proj_weight_quantizers.200 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.down_proj_weight_quantizers.201 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.97e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.down_proj_weight_quantizers.202 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.down_proj_weight_quantizers.203 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.down_proj_weight_quantizers.204 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.down_proj_weight_quantizers.205 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.down_proj_weight_quantizers.206 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.down_proj_weight_quantizers.207 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.down_proj_weight_quantizers.208 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.down_proj_weight_quantizers.209 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.down_proj_weight_quantizers.210 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.73e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.down_proj_weight_quantizers.211 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.down_proj_weight_quantizers.212 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.28e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.down_proj_weight_quantizers.213 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.down_proj_weight_quantizers.214 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.15e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.down_proj_weight_quantizers.215 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.19e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.down_proj_weight_quantizers.216 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.58e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.down_proj_weight_quantizers.217 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.down_proj_weight_quantizers.218 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.down_proj_weight_quantizers.219 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.down_proj_weight_quantizers.220 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.down_proj_weight_quantizers.221 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.30e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.down_proj_weight_quantizers.222 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.37e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.down_proj_weight_quantizers.223 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.down_proj_weight_quantizers.224 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.down_proj_weight_quantizers.225 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.15e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.down_proj_weight_quantizers.226 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.down_proj_weight_quantizers.227 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.down_proj_weight_quantizers.228 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.15e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.down_proj_weight_quantizers.229 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.down_proj_weight_quantizers.230 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.down_proj_weight_quantizers.231 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.down_proj_weight_quantizers.232 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.down_proj_weight_quantizers.233 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.down_proj_weight_quantizers.234 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.down_proj_weight_quantizers.235 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.down_proj_weight_quantizers.236 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.down_proj_weight_quantizers.237 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.down_proj_weight_quantizers.238 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.09e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.down_proj_weight_quantizers.239 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.down_proj_weight_quantizers.240 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.down_proj_weight_quantizers.241 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.80e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.down_proj_weight_quantizers.242 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.down_proj_weight_quantizers.243 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.down_proj_weight_quantizers.244 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.14e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.down_proj_weight_quantizers.245 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.down_proj_weight_quantizers.246 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.down_proj_weight_quantizers.247 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.04e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.down_proj_weight_quantizers.248 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.down_proj_weight_quantizers.249 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.18e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.down_proj_weight_quantizers.250 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.down_proj_weight_quantizers.251 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.down_proj_weight_quantizers.252 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.down_proj_weight_quantizers.253 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.down_proj_weight_quantizers.254 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.down_proj_weight_quantizers.255 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.14e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.shared_experts.gate_proj.input_quantizer TensorQuantizer(disabled) +model.layers.71.mlp.shared_experts.gate_proj.output_quantizer TensorQuantizer(disabled) +model.layers.71.mlp.shared_experts.gate_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.71.mlp.shared_experts.up_proj.input_quantizer TensorQuantizer(disabled) +model.layers.71.mlp.shared_experts.up_proj.output_quantizer TensorQuantizer(disabled) +model.layers.71.mlp.shared_experts.up_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.71.mlp.shared_experts.down_proj.input_quantizer TensorQuantizer(disabled) +model.layers.71.mlp.shared_experts.down_proj.output_quantizer TensorQuantizer(disabled) +model.layers.71.mlp.shared_experts.down_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.72.self_attn.q_a_proj.input_quantizer TensorQuantizer(disabled) +model.layers.72.self_attn.q_a_proj.output_quantizer TensorQuantizer(disabled) +model.layers.72.self_attn.q_a_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.72.self_attn.q_b_proj.input_quantizer TensorQuantizer(disabled) +model.layers.72.self_attn.q_b_proj.output_quantizer TensorQuantizer(disabled) +model.layers.72.self_attn.q_b_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.72.self_attn.kv_a_proj_with_mqa.input_quantizer TensorQuantizer(disabled) +model.layers.72.self_attn.kv_a_proj_with_mqa.output_quantizer TensorQuantizer(disabled) +model.layers.72.self_attn.kv_a_proj_with_mqa.weight_quantizer TensorQuantizer(disabled) +model.layers.72.self_attn.kv_b_proj.input_quantizer TensorQuantizer(disabled) +model.layers.72.self_attn.kv_b_proj.output_quantizer TensorQuantizer(disabled) +model.layers.72.self_attn.kv_b_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.72.self_attn.o_proj.input_quantizer TensorQuantizer(disabled) +model.layers.72.self_attn.o_proj.output_quantizer TensorQuantizer(disabled) +model.layers.72.self_attn.o_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.72.self_attn.q_bmm_quantizer TensorQuantizer(disabled) +model.layers.72.self_attn.k_bmm_quantizer TensorQuantizer((4, 3) bit fake per-tensor amax=4.48e+02(const) calibrator=MaxCalibrator quant) +model.layers.72.self_attn.v_bmm_quantizer TensorQuantizer((4, 3) bit fake per-tensor amax=4.48e+02(const) calibrator=MaxCalibrator quant) +model.layers.72.self_attn.p_bmm_quantizer TensorQuantizer(disabled) +model.layers.72.mlp.experts.gate_up_proj_input_quantizer TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.78e+00 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.gate_up_proj_weight_quantizers.0 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.gate_up_proj_weight_quantizers.1 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.31e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.gate_up_proj_weight_quantizers.2 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.gate_up_proj_weight_quantizers.3 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e+00 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.gate_up_proj_weight_quantizers.4 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.56e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.gate_up_proj_weight_quantizers.5 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.gate_up_proj_weight_quantizers.6 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.47e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.gate_up_proj_weight_quantizers.7 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.gate_up_proj_weight_quantizers.8 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.18e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.gate_up_proj_weight_quantizers.9 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.21e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.gate_up_proj_weight_quantizers.10 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=5.62e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.gate_up_proj_weight_quantizers.11 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.79e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.gate_up_proj_weight_quantizers.12 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.55e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.gate_up_proj_weight_quantizers.13 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.13e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.gate_up_proj_weight_quantizers.14 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.gate_up_proj_weight_quantizers.15 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=4.39e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.gate_up_proj_weight_quantizers.16 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.61e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.gate_up_proj_weight_quantizers.17 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.16e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.gate_up_proj_weight_quantizers.18 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.gate_up_proj_weight_quantizers.19 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.gate_up_proj_weight_quantizers.20 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.29e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.gate_up_proj_weight_quantizers.21 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.15e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.gate_up_proj_weight_quantizers.22 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.gate_up_proj_weight_quantizers.23 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.gate_up_proj_weight_quantizers.24 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.gate_up_proj_weight_quantizers.25 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.gate_up_proj_weight_quantizers.26 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.34e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.gate_up_proj_weight_quantizers.27 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.gate_up_proj_weight_quantizers.28 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.gate_up_proj_weight_quantizers.29 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=5.59e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.gate_up_proj_weight_quantizers.30 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.gate_up_proj_weight_quantizers.31 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.73e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.gate_up_proj_weight_quantizers.32 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.gate_up_proj_weight_quantizers.33 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.47e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.gate_up_proj_weight_quantizers.34 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.24e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.gate_up_proj_weight_quantizers.35 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=4.45e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.gate_up_proj_weight_quantizers.36 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.gate_up_proj_weight_quantizers.37 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.gate_up_proj_weight_quantizers.38 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.24e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.gate_up_proj_weight_quantizers.39 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.21e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.gate_up_proj_weight_quantizers.40 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.gate_up_proj_weight_quantizers.41 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.gate_up_proj_weight_quantizers.42 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.15e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.gate_up_proj_weight_quantizers.43 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.40e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.gate_up_proj_weight_quantizers.44 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.20e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.gate_up_proj_weight_quantizers.45 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.68e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.gate_up_proj_weight_quantizers.46 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.gate_up_proj_weight_quantizers.47 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.gate_up_proj_weight_quantizers.48 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.gate_up_proj_weight_quantizers.49 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.gate_up_proj_weight_quantizers.50 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.24e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.gate_up_proj_weight_quantizers.51 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.gate_up_proj_weight_quantizers.52 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.43e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.gate_up_proj_weight_quantizers.53 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.gate_up_proj_weight_quantizers.54 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.05e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.gate_up_proj_weight_quantizers.55 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.42e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.gate_up_proj_weight_quantizers.56 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.24e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.gate_up_proj_weight_quantizers.57 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.gate_up_proj_weight_quantizers.58 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.gate_up_proj_weight_quantizers.59 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.29e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.gate_up_proj_weight_quantizers.60 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=5.59e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.gate_up_proj_weight_quantizers.61 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.gate_up_proj_weight_quantizers.62 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.gate_up_proj_weight_quantizers.63 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.35e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.gate_up_proj_weight_quantizers.64 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.07e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.gate_up_proj_weight_quantizers.65 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.gate_up_proj_weight_quantizers.66 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.41e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.gate_up_proj_weight_quantizers.67 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.57e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.gate_up_proj_weight_quantizers.68 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.gate_up_proj_weight_quantizers.69 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.gate_up_proj_weight_quantizers.70 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.83e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.gate_up_proj_weight_quantizers.71 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.25e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.gate_up_proj_weight_quantizers.72 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.gate_up_proj_weight_quantizers.73 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.22e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.gate_up_proj_weight_quantizers.74 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.gate_up_proj_weight_quantizers.75 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.64e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.gate_up_proj_weight_quantizers.76 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.gate_up_proj_weight_quantizers.77 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.15e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.gate_up_proj_weight_quantizers.78 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=6.72e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.gate_up_proj_weight_quantizers.79 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.13e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.gate_up_proj_weight_quantizers.80 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.gate_up_proj_weight_quantizers.81 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.60e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.gate_up_proj_weight_quantizers.82 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.36e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.gate_up_proj_weight_quantizers.83 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.gate_up_proj_weight_quantizers.84 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.gate_up_proj_weight_quantizers.85 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.41e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.gate_up_proj_weight_quantizers.86 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.62e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.gate_up_proj_weight_quantizers.87 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.gate_up_proj_weight_quantizers.88 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.gate_up_proj_weight_quantizers.89 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.19e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.gate_up_proj_weight_quantizers.90 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.gate_up_proj_weight_quantizers.91 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.gate_up_proj_weight_quantizers.92 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.13e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.gate_up_proj_weight_quantizers.93 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.gate_up_proj_weight_quantizers.94 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.gate_up_proj_weight_quantizers.95 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.16e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.gate_up_proj_weight_quantizers.96 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.36e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.gate_up_proj_weight_quantizers.97 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.gate_up_proj_weight_quantizers.98 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.gate_up_proj_weight_quantizers.99 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.gate_up_proj_weight_quantizers.100 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.40e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.gate_up_proj_weight_quantizers.101 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.gate_up_proj_weight_quantizers.102 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.56e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.gate_up_proj_weight_quantizers.103 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.14e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.gate_up_proj_weight_quantizers.104 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.14e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.gate_up_proj_weight_quantizers.105 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.66e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.gate_up_proj_weight_quantizers.106 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.gate_up_proj_weight_quantizers.107 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.gate_up_proj_weight_quantizers.108 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.gate_up_proj_weight_quantizers.109 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.gate_up_proj_weight_quantizers.110 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.gate_up_proj_weight_quantizers.111 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.14e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.gate_up_proj_weight_quantizers.112 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.16e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.gate_up_proj_weight_quantizers.113 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.38e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.gate_up_proj_weight_quantizers.114 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.gate_up_proj_weight_quantizers.115 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.50e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.gate_up_proj_weight_quantizers.116 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.13e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.gate_up_proj_weight_quantizers.117 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.21e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.gate_up_proj_weight_quantizers.118 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.40e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.gate_up_proj_weight_quantizers.119 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=4.53e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.gate_up_proj_weight_quantizers.120 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.gate_up_proj_weight_quantizers.121 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.gate_up_proj_weight_quantizers.122 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=4.10e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.gate_up_proj_weight_quantizers.123 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.gate_up_proj_weight_quantizers.124 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.95e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.gate_up_proj_weight_quantizers.125 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.49e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.gate_up_proj_weight_quantizers.126 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.15e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.gate_up_proj_weight_quantizers.127 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.25e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.gate_up_proj_weight_quantizers.128 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.43e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.gate_up_proj_weight_quantizers.129 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.17e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.gate_up_proj_weight_quantizers.130 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.28e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.gate_up_proj_weight_quantizers.131 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.gate_up_proj_weight_quantizers.132 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.gate_up_proj_weight_quantizers.133 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.18e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.gate_up_proj_weight_quantizers.134 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.35e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.gate_up_proj_weight_quantizers.135 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.16e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.gate_up_proj_weight_quantizers.136 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.62e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.gate_up_proj_weight_quantizers.137 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.51e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.gate_up_proj_weight_quantizers.138 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.gate_up_proj_weight_quantizers.139 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.gate_up_proj_weight_quantizers.140 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.51e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.gate_up_proj_weight_quantizers.141 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.29e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.gate_up_proj_weight_quantizers.142 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.43e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.gate_up_proj_weight_quantizers.143 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.26e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.gate_up_proj_weight_quantizers.144 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.gate_up_proj_weight_quantizers.145 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.gate_up_proj_weight_quantizers.146 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.gate_up_proj_weight_quantizers.147 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.36e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.gate_up_proj_weight_quantizers.148 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.26e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.gate_up_proj_weight_quantizers.149 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.gate_up_proj_weight_quantizers.150 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.gate_up_proj_weight_quantizers.151 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.14e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.gate_up_proj_weight_quantizers.152 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.gate_up_proj_weight_quantizers.153 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.gate_up_proj_weight_quantizers.154 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.35e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.gate_up_proj_weight_quantizers.155 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.gate_up_proj_weight_quantizers.156 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.23e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.gate_up_proj_weight_quantizers.157 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.43e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.gate_up_proj_weight_quantizers.158 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.34e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.gate_up_proj_weight_quantizers.159 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.gate_up_proj_weight_quantizers.160 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.21e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.gate_up_proj_weight_quantizers.161 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.46e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.gate_up_proj_weight_quantizers.162 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.gate_up_proj_weight_quantizers.163 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.gate_up_proj_weight_quantizers.164 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.gate_up_proj_weight_quantizers.165 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.32e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.gate_up_proj_weight_quantizers.166 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.16e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.gate_up_proj_weight_quantizers.167 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.gate_up_proj_weight_quantizers.168 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.18e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.gate_up_proj_weight_quantizers.169 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.16e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.gate_up_proj_weight_quantizers.170 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.gate_up_proj_weight_quantizers.171 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.31e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.gate_up_proj_weight_quantizers.172 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.gate_up_proj_weight_quantizers.173 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.54e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.gate_up_proj_weight_quantizers.174 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.gate_up_proj_weight_quantizers.175 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.gate_up_proj_weight_quantizers.176 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.44e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.gate_up_proj_weight_quantizers.177 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.25e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.gate_up_proj_weight_quantizers.178 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.gate_up_proj_weight_quantizers.179 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.43e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.gate_up_proj_weight_quantizers.180 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.gate_up_proj_weight_quantizers.181 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.22e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.gate_up_proj_weight_quantizers.182 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.30e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.gate_up_proj_weight_quantizers.183 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.55e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.gate_up_proj_weight_quantizers.184 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.gate_up_proj_weight_quantizers.185 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.19e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.gate_up_proj_weight_quantizers.186 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.gate_up_proj_weight_quantizers.187 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.24e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.gate_up_proj_weight_quantizers.188 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.gate_up_proj_weight_quantizers.189 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.15e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.gate_up_proj_weight_quantizers.190 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.gate_up_proj_weight_quantizers.191 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.19e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.gate_up_proj_weight_quantizers.192 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.gate_up_proj_weight_quantizers.193 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.13e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.gate_up_proj_weight_quantizers.194 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.20e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.gate_up_proj_weight_quantizers.195 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.73e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.gate_up_proj_weight_quantizers.196 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.28e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.gate_up_proj_weight_quantizers.197 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.16e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.gate_up_proj_weight_quantizers.198 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.14e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.gate_up_proj_weight_quantizers.199 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.23e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.gate_up_proj_weight_quantizers.200 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.gate_up_proj_weight_quantizers.201 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.88e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.gate_up_proj_weight_quantizers.202 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.68e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.gate_up_proj_weight_quantizers.203 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.gate_up_proj_weight_quantizers.204 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.19e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.gate_up_proj_weight_quantizers.205 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.gate_up_proj_weight_quantizers.206 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.45e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.gate_up_proj_weight_quantizers.207 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.17e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.gate_up_proj_weight_quantizers.208 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.gate_up_proj_weight_quantizers.209 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.gate_up_proj_weight_quantizers.210 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.gate_up_proj_weight_quantizers.211 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.30e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.gate_up_proj_weight_quantizers.212 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.gate_up_proj_weight_quantizers.213 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.27e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.gate_up_proj_weight_quantizers.214 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.18e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.gate_up_proj_weight_quantizers.215 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.gate_up_proj_weight_quantizers.216 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.gate_up_proj_weight_quantizers.217 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.gate_up_proj_weight_quantizers.218 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.30e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.gate_up_proj_weight_quantizers.219 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.gate_up_proj_weight_quantizers.220 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.gate_up_proj_weight_quantizers.221 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.27e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.gate_up_proj_weight_quantizers.222 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.17e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.gate_up_proj_weight_quantizers.223 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.gate_up_proj_weight_quantizers.224 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.27e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.gate_up_proj_weight_quantizers.225 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.13e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.gate_up_proj_weight_quantizers.226 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.gate_up_proj_weight_quantizers.227 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.gate_up_proj_weight_quantizers.228 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.gate_up_proj_weight_quantizers.229 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.20e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.gate_up_proj_weight_quantizers.230 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.16e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.gate_up_proj_weight_quantizers.231 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.gate_up_proj_weight_quantizers.232 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.32e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.gate_up_proj_weight_quantizers.233 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.gate_up_proj_weight_quantizers.234 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.gate_up_proj_weight_quantizers.235 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.gate_up_proj_weight_quantizers.236 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.gate_up_proj_weight_quantizers.237 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.gate_up_proj_weight_quantizers.238 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.gate_up_proj_weight_quantizers.239 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.gate_up_proj_weight_quantizers.240 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.15e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.gate_up_proj_weight_quantizers.241 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.37e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.gate_up_proj_weight_quantizers.242 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.32e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.gate_up_proj_weight_quantizers.243 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.35e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.gate_up_proj_weight_quantizers.244 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.gate_up_proj_weight_quantizers.245 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.76e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.gate_up_proj_weight_quantizers.246 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.20e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.gate_up_proj_weight_quantizers.247 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.17e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.gate_up_proj_weight_quantizers.248 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.gate_up_proj_weight_quantizers.249 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.33e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.gate_up_proj_weight_quantizers.250 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.37e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.gate_up_proj_weight_quantizers.251 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.98e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.gate_up_proj_weight_quantizers.252 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.gate_up_proj_weight_quantizers.253 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.gate_up_proj_weight_quantizers.254 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.30e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.gate_up_proj_weight_quantizers.255 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.18e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.down_proj_input_quantizer TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.54e+02 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.down_proj_weight_quantizers.0 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.down_proj_weight_quantizers.1 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.17e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.down_proj_weight_quantizers.2 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.down_proj_weight_quantizers.3 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.75e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.down_proj_weight_quantizers.4 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.down_proj_weight_quantizers.5 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.down_proj_weight_quantizers.6 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.41e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.down_proj_weight_quantizers.7 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.46e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.down_proj_weight_quantizers.8 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.down_proj_weight_quantizers.9 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.down_proj_weight_quantizers.10 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.77e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.down_proj_weight_quantizers.11 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.07e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.down_proj_weight_quantizers.12 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.21e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.down_proj_weight_quantizers.13 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.down_proj_weight_quantizers.14 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.down_proj_weight_quantizers.15 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.38e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.down_proj_weight_quantizers.16 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.28e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.down_proj_weight_quantizers.17 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.15e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.down_proj_weight_quantizers.18 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.down_proj_weight_quantizers.19 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.down_proj_weight_quantizers.20 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.down_proj_weight_quantizers.21 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.down_proj_weight_quantizers.22 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.down_proj_weight_quantizers.23 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.down_proj_weight_quantizers.24 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.down_proj_weight_quantizers.25 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.down_proj_weight_quantizers.26 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.48e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.down_proj_weight_quantizers.27 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.down_proj_weight_quantizers.28 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.down_proj_weight_quantizers.29 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.68e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.down_proj_weight_quantizers.30 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.61e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.down_proj_weight_quantizers.31 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.down_proj_weight_quantizers.32 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.down_proj_weight_quantizers.33 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.20e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.down_proj_weight_quantizers.34 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.down_proj_weight_quantizers.35 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.down_proj_weight_quantizers.36 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.down_proj_weight_quantizers.37 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.down_proj_weight_quantizers.38 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.down_proj_weight_quantizers.39 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.down_proj_weight_quantizers.40 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.down_proj_weight_quantizers.41 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.down_proj_weight_quantizers.42 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.21e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.down_proj_weight_quantizers.43 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.54e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.down_proj_weight_quantizers.44 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.81e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.down_proj_weight_quantizers.45 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.35e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.down_proj_weight_quantizers.46 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.down_proj_weight_quantizers.47 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.down_proj_weight_quantizers.48 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.down_proj_weight_quantizers.49 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.down_proj_weight_quantizers.50 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.down_proj_weight_quantizers.51 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.down_proj_weight_quantizers.52 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.20e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.down_proj_weight_quantizers.53 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.down_proj_weight_quantizers.54 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.49e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.down_proj_weight_quantizers.55 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.59e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.down_proj_weight_quantizers.56 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.down_proj_weight_quantizers.57 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.down_proj_weight_quantizers.58 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.down_proj_weight_quantizers.59 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.down_proj_weight_quantizers.60 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.53e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.down_proj_weight_quantizers.61 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.down_proj_weight_quantizers.62 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.down_proj_weight_quantizers.63 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.23e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.down_proj_weight_quantizers.64 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.down_proj_weight_quantizers.65 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.down_proj_weight_quantizers.66 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.15e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.down_proj_weight_quantizers.67 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.down_proj_weight_quantizers.68 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.down_proj_weight_quantizers.69 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.down_proj_weight_quantizers.70 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=4.14e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.down_proj_weight_quantizers.71 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.down_proj_weight_quantizers.72 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.down_proj_weight_quantizers.73 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.down_proj_weight_quantizers.74 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.down_proj_weight_quantizers.75 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.down_proj_weight_quantizers.76 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.46e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.down_proj_weight_quantizers.77 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.down_proj_weight_quantizers.78 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.86e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.down_proj_weight_quantizers.79 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.down_proj_weight_quantizers.80 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.down_proj_weight_quantizers.81 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.29e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.down_proj_weight_quantizers.82 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.down_proj_weight_quantizers.83 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.down_proj_weight_quantizers.84 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.down_proj_weight_quantizers.85 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.down_proj_weight_quantizers.86 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.13e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.down_proj_weight_quantizers.87 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.65e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.down_proj_weight_quantizers.88 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.down_proj_weight_quantizers.89 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.down_proj_weight_quantizers.90 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.18e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.down_proj_weight_quantizers.91 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.down_proj_weight_quantizers.92 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.86e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.down_proj_weight_quantizers.93 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.down_proj_weight_quantizers.94 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.57e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.down_proj_weight_quantizers.95 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.down_proj_weight_quantizers.96 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.42e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.down_proj_weight_quantizers.97 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.down_proj_weight_quantizers.98 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.down_proj_weight_quantizers.99 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.down_proj_weight_quantizers.100 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.down_proj_weight_quantizers.101 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.down_proj_weight_quantizers.102 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.down_proj_weight_quantizers.103 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.down_proj_weight_quantizers.104 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.down_proj_weight_quantizers.105 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.21e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.down_proj_weight_quantizers.106 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.down_proj_weight_quantizers.107 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.down_proj_weight_quantizers.108 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.down_proj_weight_quantizers.109 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.18e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.down_proj_weight_quantizers.110 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.down_proj_weight_quantizers.111 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.down_proj_weight_quantizers.112 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.down_proj_weight_quantizers.113 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.down_proj_weight_quantizers.114 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.down_proj_weight_quantizers.115 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.down_proj_weight_quantizers.116 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.down_proj_weight_quantizers.117 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.24e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.down_proj_weight_quantizers.118 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.down_proj_weight_quantizers.119 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.25e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.down_proj_weight_quantizers.120 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.down_proj_weight_quantizers.121 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.down_proj_weight_quantizers.122 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.down_proj_weight_quantizers.123 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.down_proj_weight_quantizers.124 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.56e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.down_proj_weight_quantizers.125 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.down_proj_weight_quantizers.126 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.down_proj_weight_quantizers.127 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.down_proj_weight_quantizers.128 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.down_proj_weight_quantizers.129 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.55e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.down_proj_weight_quantizers.130 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.down_proj_weight_quantizers.131 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.62e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.down_proj_weight_quantizers.132 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.down_proj_weight_quantizers.133 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.down_proj_weight_quantizers.134 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.18e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.down_proj_weight_quantizers.135 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.down_proj_weight_quantizers.136 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=4.80e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.down_proj_weight_quantizers.137 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=4.04e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.down_proj_weight_quantizers.138 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.29e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.down_proj_weight_quantizers.139 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.down_proj_weight_quantizers.140 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.down_proj_weight_quantizers.141 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.65e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.down_proj_weight_quantizers.142 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.26e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.down_proj_weight_quantizers.143 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.down_proj_weight_quantizers.144 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.down_proj_weight_quantizers.145 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.down_proj_weight_quantizers.146 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.down_proj_weight_quantizers.147 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.28e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.down_proj_weight_quantizers.148 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.14e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.down_proj_weight_quantizers.149 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.down_proj_weight_quantizers.150 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.down_proj_weight_quantizers.151 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.39e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.down_proj_weight_quantizers.152 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.down_proj_weight_quantizers.153 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.down_proj_weight_quantizers.154 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.68e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.down_proj_weight_quantizers.155 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.down_proj_weight_quantizers.156 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.down_proj_weight_quantizers.157 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.33e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.down_proj_weight_quantizers.158 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.down_proj_weight_quantizers.159 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.down_proj_weight_quantizers.160 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.54e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.down_proj_weight_quantizers.161 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.down_proj_weight_quantizers.162 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.down_proj_weight_quantizers.163 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.down_proj_weight_quantizers.164 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.down_proj_weight_quantizers.165 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.14e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.down_proj_weight_quantizers.166 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.13e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.down_proj_weight_quantizers.167 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.down_proj_weight_quantizers.168 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.down_proj_weight_quantizers.169 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.47e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.down_proj_weight_quantizers.170 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.down_proj_weight_quantizers.171 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.49e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.down_proj_weight_quantizers.172 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.down_proj_weight_quantizers.173 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.31e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.down_proj_weight_quantizers.174 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.down_proj_weight_quantizers.175 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.down_proj_weight_quantizers.176 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.14e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.down_proj_weight_quantizers.177 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.down_proj_weight_quantizers.178 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.25e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.down_proj_weight_quantizers.179 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.down_proj_weight_quantizers.180 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.down_proj_weight_quantizers.181 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.down_proj_weight_quantizers.182 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.down_proj_weight_quantizers.183 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.79e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.down_proj_weight_quantizers.184 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.down_proj_weight_quantizers.185 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.down_proj_weight_quantizers.186 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.down_proj_weight_quantizers.187 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.down_proj_weight_quantizers.188 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.down_proj_weight_quantizers.189 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.down_proj_weight_quantizers.190 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.20e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.down_proj_weight_quantizers.191 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.down_proj_weight_quantizers.192 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.down_proj_weight_quantizers.193 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.down_proj_weight_quantizers.194 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.down_proj_weight_quantizers.195 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.down_proj_weight_quantizers.196 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.down_proj_weight_quantizers.197 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.down_proj_weight_quantizers.198 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.82e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.down_proj_weight_quantizers.199 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.down_proj_weight_quantizers.200 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.down_proj_weight_quantizers.201 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.46e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.down_proj_weight_quantizers.202 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.63e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.down_proj_weight_quantizers.203 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.down_proj_weight_quantizers.204 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.21e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.down_proj_weight_quantizers.205 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.down_proj_weight_quantizers.206 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.down_proj_weight_quantizers.207 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.down_proj_weight_quantizers.208 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.down_proj_weight_quantizers.209 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.down_proj_weight_quantizers.210 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.25e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.down_proj_weight_quantizers.211 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.down_proj_weight_quantizers.212 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.35e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.down_proj_weight_quantizers.213 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.25e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.down_proj_weight_quantizers.214 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.down_proj_weight_quantizers.215 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.down_proj_weight_quantizers.216 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.down_proj_weight_quantizers.217 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.down_proj_weight_quantizers.218 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.down_proj_weight_quantizers.219 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.down_proj_weight_quantizers.220 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.down_proj_weight_quantizers.221 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.70e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.down_proj_weight_quantizers.222 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.down_proj_weight_quantizers.223 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.down_proj_weight_quantizers.224 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.58e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.down_proj_weight_quantizers.225 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.down_proj_weight_quantizers.226 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.down_proj_weight_quantizers.227 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.down_proj_weight_quantizers.228 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.down_proj_weight_quantizers.229 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.down_proj_weight_quantizers.230 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.07e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.down_proj_weight_quantizers.231 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.17e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.down_proj_weight_quantizers.232 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.down_proj_weight_quantizers.233 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.20e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.down_proj_weight_quantizers.234 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.down_proj_weight_quantizers.235 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.down_proj_weight_quantizers.236 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.down_proj_weight_quantizers.237 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.down_proj_weight_quantizers.238 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.down_proj_weight_quantizers.239 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.down_proj_weight_quantizers.240 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.down_proj_weight_quantizers.241 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.down_proj_weight_quantizers.242 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.down_proj_weight_quantizers.243 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.down_proj_weight_quantizers.244 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.down_proj_weight_quantizers.245 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.down_proj_weight_quantizers.246 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.down_proj_weight_quantizers.247 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.down_proj_weight_quantizers.248 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.down_proj_weight_quantizers.249 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.20e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.down_proj_weight_quantizers.250 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.down_proj_weight_quantizers.251 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.16e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.down_proj_weight_quantizers.252 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.down_proj_weight_quantizers.253 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.down_proj_weight_quantizers.254 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.70e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.down_proj_weight_quantizers.255 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.shared_experts.gate_proj.input_quantizer TensorQuantizer(disabled) +model.layers.72.mlp.shared_experts.gate_proj.output_quantizer TensorQuantizer(disabled) +model.layers.72.mlp.shared_experts.gate_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.72.mlp.shared_experts.up_proj.input_quantizer TensorQuantizer(disabled) +model.layers.72.mlp.shared_experts.up_proj.output_quantizer TensorQuantizer(disabled) +model.layers.72.mlp.shared_experts.up_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.72.mlp.shared_experts.down_proj.input_quantizer TensorQuantizer(disabled) +model.layers.72.mlp.shared_experts.down_proj.output_quantizer TensorQuantizer(disabled) +model.layers.72.mlp.shared_experts.down_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.73.self_attn.q_a_proj.input_quantizer TensorQuantizer(disabled) +model.layers.73.self_attn.q_a_proj.output_quantizer TensorQuantizer(disabled) +model.layers.73.self_attn.q_a_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.73.self_attn.q_b_proj.input_quantizer TensorQuantizer(disabled) +model.layers.73.self_attn.q_b_proj.output_quantizer TensorQuantizer(disabled) +model.layers.73.self_attn.q_b_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.73.self_attn.kv_a_proj_with_mqa.input_quantizer TensorQuantizer(disabled) +model.layers.73.self_attn.kv_a_proj_with_mqa.output_quantizer TensorQuantizer(disabled) +model.layers.73.self_attn.kv_a_proj_with_mqa.weight_quantizer TensorQuantizer(disabled) +model.layers.73.self_attn.kv_b_proj.input_quantizer TensorQuantizer(disabled) +model.layers.73.self_attn.kv_b_proj.output_quantizer TensorQuantizer(disabled) +model.layers.73.self_attn.kv_b_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.73.self_attn.o_proj.input_quantizer TensorQuantizer(disabled) +model.layers.73.self_attn.o_proj.output_quantizer TensorQuantizer(disabled) +model.layers.73.self_attn.o_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.73.self_attn.q_bmm_quantizer TensorQuantizer(disabled) +model.layers.73.self_attn.k_bmm_quantizer TensorQuantizer((4, 3) bit fake per-tensor amax=4.48e+02(const) calibrator=MaxCalibrator quant) +model.layers.73.self_attn.v_bmm_quantizer TensorQuantizer((4, 3) bit fake per-tensor amax=4.48e+02(const) calibrator=MaxCalibrator quant) +model.layers.73.self_attn.p_bmm_quantizer TensorQuantizer(disabled) +model.layers.73.mlp.experts.gate_up_proj_input_quantizer TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.61e+00 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.gate_up_proj_weight_quantizers.0 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.gate_up_proj_weight_quantizers.1 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.gate_up_proj_weight_quantizers.2 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.gate_up_proj_weight_quantizers.3 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.54e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.gate_up_proj_weight_quantizers.4 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.20e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.gate_up_proj_weight_quantizers.5 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.gate_up_proj_weight_quantizers.6 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.42e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.gate_up_proj_weight_quantizers.7 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.gate_up_proj_weight_quantizers.8 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.24e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.gate_up_proj_weight_quantizers.9 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.18e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.gate_up_proj_weight_quantizers.10 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=4.63e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.gate_up_proj_weight_quantizers.11 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.17e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.gate_up_proj_weight_quantizers.12 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.gate_up_proj_weight_quantizers.13 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.38e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.gate_up_proj_weight_quantizers.14 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.31e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.gate_up_proj_weight_quantizers.15 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.gate_up_proj_weight_quantizers.16 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.gate_up_proj_weight_quantizers.17 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.gate_up_proj_weight_quantizers.18 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.gate_up_proj_weight_quantizers.19 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.48e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.gate_up_proj_weight_quantizers.20 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.gate_up_proj_weight_quantizers.21 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.34e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.gate_up_proj_weight_quantizers.22 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.37e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.gate_up_proj_weight_quantizers.23 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.gate_up_proj_weight_quantizers.24 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.gate_up_proj_weight_quantizers.25 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.34e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.gate_up_proj_weight_quantizers.26 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.15e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.gate_up_proj_weight_quantizers.27 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.gate_up_proj_weight_quantizers.28 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.gate_up_proj_weight_quantizers.29 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.63e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.gate_up_proj_weight_quantizers.30 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.47e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.gate_up_proj_weight_quantizers.31 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.09e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.gate_up_proj_weight_quantizers.32 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.24e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.gate_up_proj_weight_quantizers.33 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.48e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.gate_up_proj_weight_quantizers.34 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.gate_up_proj_weight_quantizers.35 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.14e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.gate_up_proj_weight_quantizers.36 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.gate_up_proj_weight_quantizers.37 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.gate_up_proj_weight_quantizers.38 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.gate_up_proj_weight_quantizers.39 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.gate_up_proj_weight_quantizers.40 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.28e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.gate_up_proj_weight_quantizers.41 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.25e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.gate_up_proj_weight_quantizers.42 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.24e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.gate_up_proj_weight_quantizers.43 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.gate_up_proj_weight_quantizers.44 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.gate_up_proj_weight_quantizers.45 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.27e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.gate_up_proj_weight_quantizers.46 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.gate_up_proj_weight_quantizers.47 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.gate_up_proj_weight_quantizers.48 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.18e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.gate_up_proj_weight_quantizers.49 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.gate_up_proj_weight_quantizers.50 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.gate_up_proj_weight_quantizers.51 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.gate_up_proj_weight_quantizers.52 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.gate_up_proj_weight_quantizers.53 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.gate_up_proj_weight_quantizers.54 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.gate_up_proj_weight_quantizers.55 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=5.27e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.gate_up_proj_weight_quantizers.56 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.gate_up_proj_weight_quantizers.57 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.20e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.gate_up_proj_weight_quantizers.58 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.24e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.gate_up_proj_weight_quantizers.59 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.gate_up_proj_weight_quantizers.60 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=4.28e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.gate_up_proj_weight_quantizers.61 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.05e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.gate_up_proj_weight_quantizers.62 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.gate_up_proj_weight_quantizers.63 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.43e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.gate_up_proj_weight_quantizers.64 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.25e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.gate_up_proj_weight_quantizers.65 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.gate_up_proj_weight_quantizers.66 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.26e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.gate_up_proj_weight_quantizers.67 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.95e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.gate_up_proj_weight_quantizers.68 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.gate_up_proj_weight_quantizers.69 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.79e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.gate_up_proj_weight_quantizers.70 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.32e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.gate_up_proj_weight_quantizers.71 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.gate_up_proj_weight_quantizers.72 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.59e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.gate_up_proj_weight_quantizers.73 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=4.26e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.gate_up_proj_weight_quantizers.74 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.gate_up_proj_weight_quantizers.75 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=5.16e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.gate_up_proj_weight_quantizers.76 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.34e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.gate_up_proj_weight_quantizers.77 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.gate_up_proj_weight_quantizers.78 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.13e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.gate_up_proj_weight_quantizers.79 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.49e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.gate_up_proj_weight_quantizers.80 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.41e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.gate_up_proj_weight_quantizers.81 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.31e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.gate_up_proj_weight_quantizers.82 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.gate_up_proj_weight_quantizers.83 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.15e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.gate_up_proj_weight_quantizers.84 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.22e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.gate_up_proj_weight_quantizers.85 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.gate_up_proj_weight_quantizers.86 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.12e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.gate_up_proj_weight_quantizers.87 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.18e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.gate_up_proj_weight_quantizers.88 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.gate_up_proj_weight_quantizers.89 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.gate_up_proj_weight_quantizers.90 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.35e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.gate_up_proj_weight_quantizers.91 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.46e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.gate_up_proj_weight_quantizers.92 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.75e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.gate_up_proj_weight_quantizers.93 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.gate_up_proj_weight_quantizers.94 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.46e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.gate_up_proj_weight_quantizers.95 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.gate_up_proj_weight_quantizers.96 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.gate_up_proj_weight_quantizers.97 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.gate_up_proj_weight_quantizers.98 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.62e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.gate_up_proj_weight_quantizers.99 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.36e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.gate_up_proj_weight_quantizers.100 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.46e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.gate_up_proj_weight_quantizers.101 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.23e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.gate_up_proj_weight_quantizers.102 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.31e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.gate_up_proj_weight_quantizers.103 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.17e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.gate_up_proj_weight_quantizers.104 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.89e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.gate_up_proj_weight_quantizers.105 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.gate_up_proj_weight_quantizers.106 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.gate_up_proj_weight_quantizers.107 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.47e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.gate_up_proj_weight_quantizers.108 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.39e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.gate_up_proj_weight_quantizers.109 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.gate_up_proj_weight_quantizers.110 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.gate_up_proj_weight_quantizers.111 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.21e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.gate_up_proj_weight_quantizers.112 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.gate_up_proj_weight_quantizers.113 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.22e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.gate_up_proj_weight_quantizers.114 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.gate_up_proj_weight_quantizers.115 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.25e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.gate_up_proj_weight_quantizers.116 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.50e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.gate_up_proj_weight_quantizers.117 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.gate_up_proj_weight_quantizers.118 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.17e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.gate_up_proj_weight_quantizers.119 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.gate_up_proj_weight_quantizers.120 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.18e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.gate_up_proj_weight_quantizers.121 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.27e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.gate_up_proj_weight_quantizers.122 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.gate_up_proj_weight_quantizers.123 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.gate_up_proj_weight_quantizers.124 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.13e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.gate_up_proj_weight_quantizers.125 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.21e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.gate_up_proj_weight_quantizers.126 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.87e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.gate_up_proj_weight_quantizers.127 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.gate_up_proj_weight_quantizers.128 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.15e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.gate_up_proj_weight_quantizers.129 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.56e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.gate_up_proj_weight_quantizers.130 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.34e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.gate_up_proj_weight_quantizers.131 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.gate_up_proj_weight_quantizers.132 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.19e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.gate_up_proj_weight_quantizers.133 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.18e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.gate_up_proj_weight_quantizers.134 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.77e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.gate_up_proj_weight_quantizers.135 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.27e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.gate_up_proj_weight_quantizers.136 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.gate_up_proj_weight_quantizers.137 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.46e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.gate_up_proj_weight_quantizers.138 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.gate_up_proj_weight_quantizers.139 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.43e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.gate_up_proj_weight_quantizers.140 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.46e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.gate_up_proj_weight_quantizers.141 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=4.18e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.gate_up_proj_weight_quantizers.142 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.gate_up_proj_weight_quantizers.143 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.15e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.gate_up_proj_weight_quantizers.144 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.gate_up_proj_weight_quantizers.145 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.32e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.gate_up_proj_weight_quantizers.146 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.gate_up_proj_weight_quantizers.147 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.gate_up_proj_weight_quantizers.148 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.gate_up_proj_weight_quantizers.149 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.18e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.gate_up_proj_weight_quantizers.150 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.13e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.gate_up_proj_weight_quantizers.151 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.75e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.gate_up_proj_weight_quantizers.152 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.gate_up_proj_weight_quantizers.153 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.25e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.gate_up_proj_weight_quantizers.154 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.23e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.gate_up_proj_weight_quantizers.155 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.gate_up_proj_weight_quantizers.156 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.21e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.gate_up_proj_weight_quantizers.157 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.34e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.gate_up_proj_weight_quantizers.158 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.23e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.gate_up_proj_weight_quantizers.159 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.gate_up_proj_weight_quantizers.160 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.23e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.gate_up_proj_weight_quantizers.161 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.gate_up_proj_weight_quantizers.162 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.gate_up_proj_weight_quantizers.163 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.14e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.gate_up_proj_weight_quantizers.164 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.gate_up_proj_weight_quantizers.165 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.21e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.gate_up_proj_weight_quantizers.166 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.gate_up_proj_weight_quantizers.167 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.46e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.gate_up_proj_weight_quantizers.168 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.23e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.gate_up_proj_weight_quantizers.169 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.25e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.gate_up_proj_weight_quantizers.170 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.23e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.gate_up_proj_weight_quantizers.171 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.gate_up_proj_weight_quantizers.172 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.21e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.gate_up_proj_weight_quantizers.173 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.gate_up_proj_weight_quantizers.174 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.30e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.gate_up_proj_weight_quantizers.175 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.24e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.gate_up_proj_weight_quantizers.176 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.21e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.gate_up_proj_weight_quantizers.177 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.gate_up_proj_weight_quantizers.178 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.30e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.gate_up_proj_weight_quantizers.179 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.25e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.gate_up_proj_weight_quantizers.180 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.12e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.gate_up_proj_weight_quantizers.181 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.gate_up_proj_weight_quantizers.182 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.gate_up_proj_weight_quantizers.183 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.gate_up_proj_weight_quantizers.184 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.gate_up_proj_weight_quantizers.185 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.gate_up_proj_weight_quantizers.186 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.gate_up_proj_weight_quantizers.187 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.gate_up_proj_weight_quantizers.188 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.95e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.gate_up_proj_weight_quantizers.189 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.31e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.gate_up_proj_weight_quantizers.190 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.79e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.gate_up_proj_weight_quantizers.191 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.gate_up_proj_weight_quantizers.192 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.18e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.gate_up_proj_weight_quantizers.193 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.50e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.gate_up_proj_weight_quantizers.194 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.gate_up_proj_weight_quantizers.195 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.gate_up_proj_weight_quantizers.196 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.36e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.gate_up_proj_weight_quantizers.197 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.57e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.gate_up_proj_weight_quantizers.198 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.gate_up_proj_weight_quantizers.199 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.27e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.gate_up_proj_weight_quantizers.200 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.78e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.gate_up_proj_weight_quantizers.201 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.17e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.gate_up_proj_weight_quantizers.202 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.gate_up_proj_weight_quantizers.203 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.gate_up_proj_weight_quantizers.204 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.gate_up_proj_weight_quantizers.205 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.gate_up_proj_weight_quantizers.206 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=4.47e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.gate_up_proj_weight_quantizers.207 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.gate_up_proj_weight_quantizers.208 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.40e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.gate_up_proj_weight_quantizers.209 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.gate_up_proj_weight_quantizers.210 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.gate_up_proj_weight_quantizers.211 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.11e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.gate_up_proj_weight_quantizers.212 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.25e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.gate_up_proj_weight_quantizers.213 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.gate_up_proj_weight_quantizers.214 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.91e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.gate_up_proj_weight_quantizers.215 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.14e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.gate_up_proj_weight_quantizers.216 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.19e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.gate_up_proj_weight_quantizers.217 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=4.08e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.gate_up_proj_weight_quantizers.218 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.gate_up_proj_weight_quantizers.219 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.gate_up_proj_weight_quantizers.220 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.65e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.gate_up_proj_weight_quantizers.221 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.gate_up_proj_weight_quantizers.222 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.gate_up_proj_weight_quantizers.223 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.05e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.gate_up_proj_weight_quantizers.224 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.38e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.gate_up_proj_weight_quantizers.225 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=6.09e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.gate_up_proj_weight_quantizers.226 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.21e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.gate_up_proj_weight_quantizers.227 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.gate_up_proj_weight_quantizers.228 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.gate_up_proj_weight_quantizers.229 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.gate_up_proj_weight_quantizers.230 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.gate_up_proj_weight_quantizers.231 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.39e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.gate_up_proj_weight_quantizers.232 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.23e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.gate_up_proj_weight_quantizers.233 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.14e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.gate_up_proj_weight_quantizers.234 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.18e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.gate_up_proj_weight_quantizers.235 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.gate_up_proj_weight_quantizers.236 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.14e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.gate_up_proj_weight_quantizers.237 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.77e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.gate_up_proj_weight_quantizers.238 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.91e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.gate_up_proj_weight_quantizers.239 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.20e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.gate_up_proj_weight_quantizers.240 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.28e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.gate_up_proj_weight_quantizers.241 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.19e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.gate_up_proj_weight_quantizers.242 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.gate_up_proj_weight_quantizers.243 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.57e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.gate_up_proj_weight_quantizers.244 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.gate_up_proj_weight_quantizers.245 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.28e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.gate_up_proj_weight_quantizers.246 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.24e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.gate_up_proj_weight_quantizers.247 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.gate_up_proj_weight_quantizers.248 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.gate_up_proj_weight_quantizers.249 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.gate_up_proj_weight_quantizers.250 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.53e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.gate_up_proj_weight_quantizers.251 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.19e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.gate_up_proj_weight_quantizers.252 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.gate_up_proj_weight_quantizers.253 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.15e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.gate_up_proj_weight_quantizers.254 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.24e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.gate_up_proj_weight_quantizers.255 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.down_proj_input_quantizer TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.50e+02 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.down_proj_weight_quantizers.0 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.down_proj_weight_quantizers.1 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.down_proj_weight_quantizers.2 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.down_proj_weight_quantizers.3 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.08e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.down_proj_weight_quantizers.4 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.down_proj_weight_quantizers.5 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.down_proj_weight_quantizers.6 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.88e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.down_proj_weight_quantizers.7 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.down_proj_weight_quantizers.8 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.down_proj_weight_quantizers.9 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.down_proj_weight_quantizers.10 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.57e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.down_proj_weight_quantizers.11 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.down_proj_weight_quantizers.12 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.down_proj_weight_quantizers.13 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.30e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.down_proj_weight_quantizers.14 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.down_proj_weight_quantizers.15 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.down_proj_weight_quantizers.16 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.down_proj_weight_quantizers.17 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.down_proj_weight_quantizers.18 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.down_proj_weight_quantizers.19 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.down_proj_weight_quantizers.20 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.down_proj_weight_quantizers.21 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.84e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.down_proj_weight_quantizers.22 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.down_proj_weight_quantizers.23 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.down_proj_weight_quantizers.24 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.down_proj_weight_quantizers.25 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.38e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.down_proj_weight_quantizers.26 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.down_proj_weight_quantizers.27 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.down_proj_weight_quantizers.28 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.down_proj_weight_quantizers.29 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.02e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.down_proj_weight_quantizers.30 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.55e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.down_proj_weight_quantizers.31 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.down_proj_weight_quantizers.32 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.16e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.down_proj_weight_quantizers.33 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.54e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.down_proj_weight_quantizers.34 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.down_proj_weight_quantizers.35 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.down_proj_weight_quantizers.36 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.17e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.down_proj_weight_quantizers.37 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.36e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.down_proj_weight_quantizers.38 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.down_proj_weight_quantizers.39 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.down_proj_weight_quantizers.40 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.down_proj_weight_quantizers.41 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.down_proj_weight_quantizers.42 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.down_proj_weight_quantizers.43 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.down_proj_weight_quantizers.44 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.26e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.down_proj_weight_quantizers.45 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.down_proj_weight_quantizers.46 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.13e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.down_proj_weight_quantizers.47 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.36e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.down_proj_weight_quantizers.48 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.down_proj_weight_quantizers.49 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.down_proj_weight_quantizers.50 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.down_proj_weight_quantizers.51 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.down_proj_weight_quantizers.52 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.down_proj_weight_quantizers.53 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.down_proj_weight_quantizers.54 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.down_proj_weight_quantizers.55 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.87e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.down_proj_weight_quantizers.56 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.down_proj_weight_quantizers.57 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.down_proj_weight_quantizers.58 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.down_proj_weight_quantizers.59 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.down_proj_weight_quantizers.60 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.27e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.down_proj_weight_quantizers.61 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.50e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.down_proj_weight_quantizers.62 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.down_proj_weight_quantizers.63 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.down_proj_weight_quantizers.64 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.down_proj_weight_quantizers.65 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.down_proj_weight_quantizers.66 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.down_proj_weight_quantizers.67 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.54e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.down_proj_weight_quantizers.68 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.down_proj_weight_quantizers.69 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.98e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.down_proj_weight_quantizers.70 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.down_proj_weight_quantizers.71 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.down_proj_weight_quantizers.72 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.15e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.down_proj_weight_quantizers.73 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.61e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.down_proj_weight_quantizers.74 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.down_proj_weight_quantizers.75 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.84e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.down_proj_weight_quantizers.76 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.down_proj_weight_quantizers.77 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.down_proj_weight_quantizers.78 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.down_proj_weight_quantizers.79 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.down_proj_weight_quantizers.80 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.00e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.down_proj_weight_quantizers.81 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.down_proj_weight_quantizers.82 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.08e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.down_proj_weight_quantizers.83 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.down_proj_weight_quantizers.84 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.down_proj_weight_quantizers.85 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.down_proj_weight_quantizers.86 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=5.16e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.down_proj_weight_quantizers.87 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.down_proj_weight_quantizers.88 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.down_proj_weight_quantizers.89 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.down_proj_weight_quantizers.90 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.down_proj_weight_quantizers.91 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.63e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.down_proj_weight_quantizers.92 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.61e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.down_proj_weight_quantizers.93 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.down_proj_weight_quantizers.94 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.91e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.down_proj_weight_quantizers.95 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.down_proj_weight_quantizers.96 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.down_proj_weight_quantizers.97 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.69e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.down_proj_weight_quantizers.98 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.down_proj_weight_quantizers.99 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.down_proj_weight_quantizers.100 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.16e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.down_proj_weight_quantizers.101 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.down_proj_weight_quantizers.102 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.60e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.down_proj_weight_quantizers.103 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.down_proj_weight_quantizers.104 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.down_proj_weight_quantizers.105 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.down_proj_weight_quantizers.106 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.down_proj_weight_quantizers.107 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.down_proj_weight_quantizers.108 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.down_proj_weight_quantizers.109 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.down_proj_weight_quantizers.110 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.down_proj_weight_quantizers.111 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.down_proj_weight_quantizers.112 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.down_proj_weight_quantizers.113 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.50e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.down_proj_weight_quantizers.114 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.down_proj_weight_quantizers.115 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.24e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.down_proj_weight_quantizers.116 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.22e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.down_proj_weight_quantizers.117 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.17e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.down_proj_weight_quantizers.118 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.16e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.down_proj_weight_quantizers.119 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.down_proj_weight_quantizers.120 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.22e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.down_proj_weight_quantizers.121 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.down_proj_weight_quantizers.122 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.down_proj_weight_quantizers.123 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.down_proj_weight_quantizers.124 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.down_proj_weight_quantizers.125 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.down_proj_weight_quantizers.126 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.77e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.down_proj_weight_quantizers.127 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.down_proj_weight_quantizers.128 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.down_proj_weight_quantizers.129 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.down_proj_weight_quantizers.130 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.down_proj_weight_quantizers.131 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.down_proj_weight_quantizers.132 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.down_proj_weight_quantizers.133 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.41e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.down_proj_weight_quantizers.134 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.96e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.down_proj_weight_quantizers.135 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.down_proj_weight_quantizers.136 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.down_proj_weight_quantizers.137 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.57e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.down_proj_weight_quantizers.138 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.down_proj_weight_quantizers.139 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.47e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.down_proj_weight_quantizers.140 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.27e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.down_proj_weight_quantizers.141 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.down_proj_weight_quantizers.142 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.down_proj_weight_quantizers.143 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.down_proj_weight_quantizers.144 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.down_proj_weight_quantizers.145 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.down_proj_weight_quantizers.146 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.down_proj_weight_quantizers.147 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.down_proj_weight_quantizers.148 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.down_proj_weight_quantizers.149 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.down_proj_weight_quantizers.150 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.down_proj_weight_quantizers.151 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.42e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.down_proj_weight_quantizers.152 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.40e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.down_proj_weight_quantizers.153 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.73e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.down_proj_weight_quantizers.154 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.down_proj_weight_quantizers.155 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.down_proj_weight_quantizers.156 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.down_proj_weight_quantizers.157 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.42e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.down_proj_weight_quantizers.158 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.down_proj_weight_quantizers.159 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.down_proj_weight_quantizers.160 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.19e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.down_proj_weight_quantizers.161 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.down_proj_weight_quantizers.162 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.13e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.down_proj_weight_quantizers.163 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.down_proj_weight_quantizers.164 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.down_proj_weight_quantizers.165 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.down_proj_weight_quantizers.166 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.down_proj_weight_quantizers.167 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.14e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.down_proj_weight_quantizers.168 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.down_proj_weight_quantizers.169 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.down_proj_weight_quantizers.170 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.down_proj_weight_quantizers.171 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.68e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.down_proj_weight_quantizers.172 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.down_proj_weight_quantizers.173 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.down_proj_weight_quantizers.174 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.64e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.down_proj_weight_quantizers.175 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.down_proj_weight_quantizers.176 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.down_proj_weight_quantizers.177 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.down_proj_weight_quantizers.178 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.34e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.down_proj_weight_quantizers.179 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.down_proj_weight_quantizers.180 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.89e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.down_proj_weight_quantizers.181 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.down_proj_weight_quantizers.182 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.down_proj_weight_quantizers.183 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.down_proj_weight_quantizers.184 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.down_proj_weight_quantizers.185 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.down_proj_weight_quantizers.186 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.down_proj_weight_quantizers.187 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.down_proj_weight_quantizers.188 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.27e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.down_proj_weight_quantizers.189 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.down_proj_weight_quantizers.190 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.20e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.down_proj_weight_quantizers.191 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.down_proj_weight_quantizers.192 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.16e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.down_proj_weight_quantizers.193 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.14e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.down_proj_weight_quantizers.194 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.53e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.down_proj_weight_quantizers.195 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.down_proj_weight_quantizers.196 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.14e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.down_proj_weight_quantizers.197 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.46e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.down_proj_weight_quantizers.198 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.42e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.down_proj_weight_quantizers.199 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.down_proj_weight_quantizers.200 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.down_proj_weight_quantizers.201 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.24e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.down_proj_weight_quantizers.202 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.33e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.down_proj_weight_quantizers.203 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.down_proj_weight_quantizers.204 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.down_proj_weight_quantizers.205 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.down_proj_weight_quantizers.206 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.14e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.down_proj_weight_quantizers.207 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.down_proj_weight_quantizers.208 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.73e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.down_proj_weight_quantizers.209 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.down_proj_weight_quantizers.210 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.down_proj_weight_quantizers.211 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.75e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.down_proj_weight_quantizers.212 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.down_proj_weight_quantizers.213 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.down_proj_weight_quantizers.214 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.36e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.down_proj_weight_quantizers.215 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.down_proj_weight_quantizers.216 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.71e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.down_proj_weight_quantizers.217 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.81e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.down_proj_weight_quantizers.218 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.down_proj_weight_quantizers.219 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.31e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.down_proj_weight_quantizers.220 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.46e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.down_proj_weight_quantizers.221 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.down_proj_weight_quantizers.222 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.down_proj_weight_quantizers.223 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.77e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.down_proj_weight_quantizers.224 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.85e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.down_proj_weight_quantizers.225 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.40e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.down_proj_weight_quantizers.226 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.40e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.down_proj_weight_quantizers.227 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.down_proj_weight_quantizers.228 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.down_proj_weight_quantizers.229 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.down_proj_weight_quantizers.230 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.down_proj_weight_quantizers.231 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.76e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.down_proj_weight_quantizers.232 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.13e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.down_proj_weight_quantizers.233 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.down_proj_weight_quantizers.234 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.35e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.down_proj_weight_quantizers.235 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.down_proj_weight_quantizers.236 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.down_proj_weight_quantizers.237 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.09e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.down_proj_weight_quantizers.238 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.79e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.down_proj_weight_quantizers.239 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.down_proj_weight_quantizers.240 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.62e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.down_proj_weight_quantizers.241 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.down_proj_weight_quantizers.242 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.down_proj_weight_quantizers.243 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.02e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.down_proj_weight_quantizers.244 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.down_proj_weight_quantizers.245 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.32e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.down_proj_weight_quantizers.246 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.down_proj_weight_quantizers.247 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.down_proj_weight_quantizers.248 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.down_proj_weight_quantizers.249 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.down_proj_weight_quantizers.250 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.down_proj_weight_quantizers.251 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.down_proj_weight_quantizers.252 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.down_proj_weight_quantizers.253 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.50e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.down_proj_weight_quantizers.254 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.30e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.down_proj_weight_quantizers.255 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.23e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.shared_experts.gate_proj.input_quantizer TensorQuantizer(disabled) +model.layers.73.mlp.shared_experts.gate_proj.output_quantizer TensorQuantizer(disabled) +model.layers.73.mlp.shared_experts.gate_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.73.mlp.shared_experts.up_proj.input_quantizer TensorQuantizer(disabled) +model.layers.73.mlp.shared_experts.up_proj.output_quantizer TensorQuantizer(disabled) +model.layers.73.mlp.shared_experts.up_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.73.mlp.shared_experts.down_proj.input_quantizer TensorQuantizer(disabled) +model.layers.73.mlp.shared_experts.down_proj.output_quantizer TensorQuantizer(disabled) +model.layers.73.mlp.shared_experts.down_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.74.self_attn.q_a_proj.input_quantizer TensorQuantizer(disabled) +model.layers.74.self_attn.q_a_proj.output_quantizer TensorQuantizer(disabled) +model.layers.74.self_attn.q_a_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.74.self_attn.q_b_proj.input_quantizer TensorQuantizer(disabled) +model.layers.74.self_attn.q_b_proj.output_quantizer TensorQuantizer(disabled) +model.layers.74.self_attn.q_b_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.74.self_attn.kv_a_proj_with_mqa.input_quantizer TensorQuantizer(disabled) +model.layers.74.self_attn.kv_a_proj_with_mqa.output_quantizer TensorQuantizer(disabled) +model.layers.74.self_attn.kv_a_proj_with_mqa.weight_quantizer TensorQuantizer(disabled) +model.layers.74.self_attn.kv_b_proj.input_quantizer TensorQuantizer(disabled) +model.layers.74.self_attn.kv_b_proj.output_quantizer TensorQuantizer(disabled) +model.layers.74.self_attn.kv_b_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.74.self_attn.o_proj.input_quantizer TensorQuantizer(disabled) +model.layers.74.self_attn.o_proj.output_quantizer TensorQuantizer(disabled) +model.layers.74.self_attn.o_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.74.self_attn.indexer.wq_b.input_quantizer TensorQuantizer(disabled) +model.layers.74.self_attn.indexer.wq_b.output_quantizer TensorQuantizer(disabled) +model.layers.74.self_attn.indexer.wq_b.weight_quantizer TensorQuantizer(disabled) +model.layers.74.self_attn.indexer.wk.input_quantizer TensorQuantizer(disabled) +model.layers.74.self_attn.indexer.wk.output_quantizer TensorQuantizer(disabled) +model.layers.74.self_attn.indexer.wk.weight_quantizer TensorQuantizer(disabled) +model.layers.74.self_attn.indexer.k_norm.input_quantizer TensorQuantizer(disabled) +model.layers.74.self_attn.indexer.k_norm.output_quantizer TensorQuantizer(disabled) +model.layers.74.self_attn.indexer.weights_proj.input_quantizer TensorQuantizer(disabled) +model.layers.74.self_attn.indexer.weights_proj.output_quantizer TensorQuantizer(disabled) +model.layers.74.self_attn.indexer.weights_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.74.self_attn.q_bmm_quantizer TensorQuantizer(disabled) +model.layers.74.self_attn.k_bmm_quantizer TensorQuantizer((4, 3) bit fake per-tensor amax=4.48e+02(const) calibrator=MaxCalibrator quant) +model.layers.74.self_attn.v_bmm_quantizer TensorQuantizer((4, 3) bit fake per-tensor amax=4.48e+02(const) calibrator=MaxCalibrator quant) +model.layers.74.self_attn.p_bmm_quantizer TensorQuantizer(disabled) +model.layers.74.mlp.experts.gate_up_proj_input_quantizer TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.88e+00 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.gate_up_proj_weight_quantizers.0 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.gate_up_proj_weight_quantizers.1 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.gate_up_proj_weight_quantizers.2 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.45e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.gate_up_proj_weight_quantizers.3 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.19e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.gate_up_proj_weight_quantizers.4 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.gate_up_proj_weight_quantizers.5 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.29e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.gate_up_proj_weight_quantizers.6 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.gate_up_proj_weight_quantizers.7 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.gate_up_proj_weight_quantizers.8 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.25e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.gate_up_proj_weight_quantizers.9 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.62e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.gate_up_proj_weight_quantizers.10 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.gate_up_proj_weight_quantizers.11 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.gate_up_proj_weight_quantizers.12 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.gate_up_proj_weight_quantizers.13 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.23e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.gate_up_proj_weight_quantizers.14 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.gate_up_proj_weight_quantizers.15 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.87e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.gate_up_proj_weight_quantizers.16 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.03e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.gate_up_proj_weight_quantizers.17 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.gate_up_proj_weight_quantizers.18 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.26e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.gate_up_proj_weight_quantizers.19 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.gate_up_proj_weight_quantizers.20 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.19e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.gate_up_proj_weight_quantizers.21 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.gate_up_proj_weight_quantizers.22 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.29e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.gate_up_proj_weight_quantizers.23 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.96e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.gate_up_proj_weight_quantizers.24 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.79e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.gate_up_proj_weight_quantizers.25 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.64e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.gate_up_proj_weight_quantizers.26 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.95e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.gate_up_proj_weight_quantizers.27 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.26e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.gate_up_proj_weight_quantizers.28 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.20e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.gate_up_proj_weight_quantizers.29 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.gate_up_proj_weight_quantizers.30 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.gate_up_proj_weight_quantizers.31 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.37e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.gate_up_proj_weight_quantizers.32 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.20e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.gate_up_proj_weight_quantizers.33 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.32e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.gate_up_proj_weight_quantizers.34 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.14e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.gate_up_proj_weight_quantizers.35 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.gate_up_proj_weight_quantizers.36 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.25e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.gate_up_proj_weight_quantizers.37 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.gate_up_proj_weight_quantizers.38 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.gate_up_proj_weight_quantizers.39 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.gate_up_proj_weight_quantizers.40 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.gate_up_proj_weight_quantizers.41 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.gate_up_proj_weight_quantizers.42 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.36e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.gate_up_proj_weight_quantizers.43 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.41e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.gate_up_proj_weight_quantizers.44 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.26e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.gate_up_proj_weight_quantizers.45 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.gate_up_proj_weight_quantizers.46 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.33e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.gate_up_proj_weight_quantizers.47 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.gate_up_proj_weight_quantizers.48 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.gate_up_proj_weight_quantizers.49 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.22e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.gate_up_proj_weight_quantizers.50 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.gate_up_proj_weight_quantizers.51 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.gate_up_proj_weight_quantizers.52 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.23e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.gate_up_proj_weight_quantizers.53 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.27e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.gate_up_proj_weight_quantizers.54 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.gate_up_proj_weight_quantizers.55 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.13e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.gate_up_proj_weight_quantizers.56 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.31e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.gate_up_proj_weight_quantizers.57 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.89e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.gate_up_proj_weight_quantizers.58 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.gate_up_proj_weight_quantizers.59 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.19e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.gate_up_proj_weight_quantizers.60 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.55e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.gate_up_proj_weight_quantizers.61 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=4.18e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.gate_up_proj_weight_quantizers.62 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.gate_up_proj_weight_quantizers.63 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.gate_up_proj_weight_quantizers.64 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.gate_up_proj_weight_quantizers.65 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.gate_up_proj_weight_quantizers.66 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.gate_up_proj_weight_quantizers.67 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.gate_up_proj_weight_quantizers.68 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.gate_up_proj_weight_quantizers.69 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.22e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.gate_up_proj_weight_quantizers.70 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.74e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.gate_up_proj_weight_quantizers.71 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.gate_up_proj_weight_quantizers.72 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.34e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.gate_up_proj_weight_quantizers.73 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.gate_up_proj_weight_quantizers.74 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.13e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.gate_up_proj_weight_quantizers.75 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.16e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.gate_up_proj_weight_quantizers.76 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.gate_up_proj_weight_quantizers.77 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.16e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.gate_up_proj_weight_quantizers.78 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.20e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.gate_up_proj_weight_quantizers.79 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.gate_up_proj_weight_quantizers.80 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.gate_up_proj_weight_quantizers.81 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.32e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.gate_up_proj_weight_quantizers.82 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.28e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.gate_up_proj_weight_quantizers.83 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.gate_up_proj_weight_quantizers.84 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.gate_up_proj_weight_quantizers.85 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.21e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.gate_up_proj_weight_quantizers.86 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.gate_up_proj_weight_quantizers.87 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.gate_up_proj_weight_quantizers.88 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.gate_up_proj_weight_quantizers.89 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.gate_up_proj_weight_quantizers.90 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.gate_up_proj_weight_quantizers.91 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.gate_up_proj_weight_quantizers.92 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.39e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.gate_up_proj_weight_quantizers.93 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.15e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.gate_up_proj_weight_quantizers.94 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.gate_up_proj_weight_quantizers.95 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.22e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.gate_up_proj_weight_quantizers.96 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.21e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.gate_up_proj_weight_quantizers.97 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.gate_up_proj_weight_quantizers.98 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.gate_up_proj_weight_quantizers.99 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.27e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.gate_up_proj_weight_quantizers.100 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.gate_up_proj_weight_quantizers.101 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.gate_up_proj_weight_quantizers.102 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=4.59e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.gate_up_proj_weight_quantizers.103 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.gate_up_proj_weight_quantizers.104 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.gate_up_proj_weight_quantizers.105 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.gate_up_proj_weight_quantizers.106 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.44e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.gate_up_proj_weight_quantizers.107 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.16e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.gate_up_proj_weight_quantizers.108 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.gate_up_proj_weight_quantizers.109 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.gate_up_proj_weight_quantizers.110 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.17e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.gate_up_proj_weight_quantizers.111 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.17e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.gate_up_proj_weight_quantizers.112 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.33e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.gate_up_proj_weight_quantizers.113 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.15e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.gate_up_proj_weight_quantizers.114 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.34e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.gate_up_proj_weight_quantizers.115 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.19e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.gate_up_proj_weight_quantizers.116 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.08e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.gate_up_proj_weight_quantizers.117 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.gate_up_proj_weight_quantizers.118 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.18e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.gate_up_proj_weight_quantizers.119 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.17e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.gate_up_proj_weight_quantizers.120 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.gate_up_proj_weight_quantizers.121 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.27e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.gate_up_proj_weight_quantizers.122 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.34e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.gate_up_proj_weight_quantizers.123 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.27e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.gate_up_proj_weight_quantizers.124 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.gate_up_proj_weight_quantizers.125 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.27e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.gate_up_proj_weight_quantizers.126 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.14e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.gate_up_proj_weight_quantizers.127 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.gate_up_proj_weight_quantizers.128 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.27e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.gate_up_proj_weight_quantizers.129 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.34e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.gate_up_proj_weight_quantizers.130 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.16e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.gate_up_proj_weight_quantizers.131 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.gate_up_proj_weight_quantizers.132 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.13e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.gate_up_proj_weight_quantizers.133 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.gate_up_proj_weight_quantizers.134 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.gate_up_proj_weight_quantizers.135 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.gate_up_proj_weight_quantizers.136 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.24e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.gate_up_proj_weight_quantizers.137 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.59e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.gate_up_proj_weight_quantizers.138 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.gate_up_proj_weight_quantizers.139 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.gate_up_proj_weight_quantizers.140 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.gate_up_proj_weight_quantizers.141 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.48e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.gate_up_proj_weight_quantizers.142 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.19e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.gate_up_proj_weight_quantizers.143 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.gate_up_proj_weight_quantizers.144 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.45e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.gate_up_proj_weight_quantizers.145 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.gate_up_proj_weight_quantizers.146 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.gate_up_proj_weight_quantizers.147 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.11e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.gate_up_proj_weight_quantizers.148 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.15e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.gate_up_proj_weight_quantizers.149 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.41e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.gate_up_proj_weight_quantizers.150 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.31e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.gate_up_proj_weight_quantizers.151 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.59e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.gate_up_proj_weight_quantizers.152 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.gate_up_proj_weight_quantizers.153 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.gate_up_proj_weight_quantizers.154 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.gate_up_proj_weight_quantizers.155 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.34e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.gate_up_proj_weight_quantizers.156 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.27e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.gate_up_proj_weight_quantizers.157 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.gate_up_proj_weight_quantizers.158 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.13e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.gate_up_proj_weight_quantizers.159 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.18e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.gate_up_proj_weight_quantizers.160 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.gate_up_proj_weight_quantizers.161 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.51e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.gate_up_proj_weight_quantizers.162 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.14e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.gate_up_proj_weight_quantizers.163 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.gate_up_proj_weight_quantizers.164 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.17e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.gate_up_proj_weight_quantizers.165 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.18e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.gate_up_proj_weight_quantizers.166 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.gate_up_proj_weight_quantizers.167 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.36e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.gate_up_proj_weight_quantizers.168 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.gate_up_proj_weight_quantizers.169 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.13e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.gate_up_proj_weight_quantizers.170 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.13e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.gate_up_proj_weight_quantizers.171 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.39e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.gate_up_proj_weight_quantizers.172 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.57e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.gate_up_proj_weight_quantizers.173 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.gate_up_proj_weight_quantizers.174 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.49e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.gate_up_proj_weight_quantizers.175 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.31e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.gate_up_proj_weight_quantizers.176 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.gate_up_proj_weight_quantizers.177 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.gate_up_proj_weight_quantizers.178 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.20e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.gate_up_proj_weight_quantizers.179 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.gate_up_proj_weight_quantizers.180 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.gate_up_proj_weight_quantizers.181 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.25e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.gate_up_proj_weight_quantizers.182 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.00e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.gate_up_proj_weight_quantizers.183 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.27e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.gate_up_proj_weight_quantizers.184 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.26e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.gate_up_proj_weight_quantizers.185 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.gate_up_proj_weight_quantizers.186 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.gate_up_proj_weight_quantizers.187 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.91e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.gate_up_proj_weight_quantizers.188 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.gate_up_proj_weight_quantizers.189 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.25e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.gate_up_proj_weight_quantizers.190 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.72e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.gate_up_proj_weight_quantizers.191 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.17e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.gate_up_proj_weight_quantizers.192 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.75e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.gate_up_proj_weight_quantizers.193 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.gate_up_proj_weight_quantizers.194 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.gate_up_proj_weight_quantizers.195 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.38e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.gate_up_proj_weight_quantizers.196 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.gate_up_proj_weight_quantizers.197 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.gate_up_proj_weight_quantizers.198 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.54e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.gate_up_proj_weight_quantizers.199 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.15e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.gate_up_proj_weight_quantizers.200 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.26e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.gate_up_proj_weight_quantizers.201 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.21e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.gate_up_proj_weight_quantizers.202 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.gate_up_proj_weight_quantizers.203 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.gate_up_proj_weight_quantizers.204 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.gate_up_proj_weight_quantizers.205 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.32e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.gate_up_proj_weight_quantizers.206 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.18e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.gate_up_proj_weight_quantizers.207 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.32e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.gate_up_proj_weight_quantizers.208 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.gate_up_proj_weight_quantizers.209 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.gate_up_proj_weight_quantizers.210 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.gate_up_proj_weight_quantizers.211 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.23e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.gate_up_proj_weight_quantizers.212 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.gate_up_proj_weight_quantizers.213 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.22e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.gate_up_proj_weight_quantizers.214 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.60e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.gate_up_proj_weight_quantizers.215 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.gate_up_proj_weight_quantizers.216 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.gate_up_proj_weight_quantizers.217 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.25e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.gate_up_proj_weight_quantizers.218 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.gate_up_proj_weight_quantizers.219 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.27e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.gate_up_proj_weight_quantizers.220 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.gate_up_proj_weight_quantizers.221 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.gate_up_proj_weight_quantizers.222 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.13e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.gate_up_proj_weight_quantizers.223 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.gate_up_proj_weight_quantizers.224 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.gate_up_proj_weight_quantizers.225 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.gate_up_proj_weight_quantizers.226 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.20e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.gate_up_proj_weight_quantizers.227 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.gate_up_proj_weight_quantizers.228 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.gate_up_proj_weight_quantizers.229 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.19e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.gate_up_proj_weight_quantizers.230 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.gate_up_proj_weight_quantizers.231 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.20e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.gate_up_proj_weight_quantizers.232 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.34e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.gate_up_proj_weight_quantizers.233 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.17e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.gate_up_proj_weight_quantizers.234 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.15e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.gate_up_proj_weight_quantizers.235 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.34e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.gate_up_proj_weight_quantizers.236 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.gate_up_proj_weight_quantizers.237 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.19e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.gate_up_proj_weight_quantizers.238 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.57e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.gate_up_proj_weight_quantizers.239 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.gate_up_proj_weight_quantizers.240 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.gate_up_proj_weight_quantizers.241 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.gate_up_proj_weight_quantizers.242 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.27e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.gate_up_proj_weight_quantizers.243 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.51e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.gate_up_proj_weight_quantizers.244 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.32e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.gate_up_proj_weight_quantizers.245 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.gate_up_proj_weight_quantizers.246 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.16e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.gate_up_proj_weight_quantizers.247 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.55e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.gate_up_proj_weight_quantizers.248 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.31e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.gate_up_proj_weight_quantizers.249 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.70e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.gate_up_proj_weight_quantizers.250 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.30e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.gate_up_proj_weight_quantizers.251 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.93e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.gate_up_proj_weight_quantizers.252 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.gate_up_proj_weight_quantizers.253 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.gate_up_proj_weight_quantizers.254 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.gate_up_proj_weight_quantizers.255 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.down_proj_input_quantizer TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.29e+02 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.down_proj_weight_quantizers.0 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.down_proj_weight_quantizers.1 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.39e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.down_proj_weight_quantizers.2 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.62e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.down_proj_weight_quantizers.3 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.14e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.down_proj_weight_quantizers.4 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.down_proj_weight_quantizers.5 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.19e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.down_proj_weight_quantizers.6 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.down_proj_weight_quantizers.7 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.down_proj_weight_quantizers.8 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.down_proj_weight_quantizers.9 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.96e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.down_proj_weight_quantizers.10 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.down_proj_weight_quantizers.11 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.down_proj_weight_quantizers.12 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.down_proj_weight_quantizers.13 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.down_proj_weight_quantizers.14 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.down_proj_weight_quantizers.15 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.32e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.down_proj_weight_quantizers.16 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.87e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.down_proj_weight_quantizers.17 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.down_proj_weight_quantizers.18 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.down_proj_weight_quantizers.19 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.down_proj_weight_quantizers.20 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.down_proj_weight_quantizers.21 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.down_proj_weight_quantizers.22 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.17e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.down_proj_weight_quantizers.23 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.44e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.down_proj_weight_quantizers.24 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.48e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.down_proj_weight_quantizers.25 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.down_proj_weight_quantizers.26 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.27e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.down_proj_weight_quantizers.27 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.32e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.down_proj_weight_quantizers.28 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.79e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.down_proj_weight_quantizers.29 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.down_proj_weight_quantizers.30 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.56e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.down_proj_weight_quantizers.31 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.down_proj_weight_quantizers.32 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.down_proj_weight_quantizers.33 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.down_proj_weight_quantizers.34 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.10e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.down_proj_weight_quantizers.35 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.down_proj_weight_quantizers.36 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.down_proj_weight_quantizers.37 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.down_proj_weight_quantizers.38 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.down_proj_weight_quantizers.39 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.down_proj_weight_quantizers.40 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.down_proj_weight_quantizers.41 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.down_proj_weight_quantizers.42 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.13e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.down_proj_weight_quantizers.43 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.down_proj_weight_quantizers.44 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.down_proj_weight_quantizers.45 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.down_proj_weight_quantizers.46 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.47e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.down_proj_weight_quantizers.47 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.down_proj_weight_quantizers.48 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.16e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.down_proj_weight_quantizers.49 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.down_proj_weight_quantizers.50 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.down_proj_weight_quantizers.51 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.down_proj_weight_quantizers.52 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.44e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.down_proj_weight_quantizers.53 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.98e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.down_proj_weight_quantizers.54 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.down_proj_weight_quantizers.55 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.down_proj_weight_quantizers.56 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.down_proj_weight_quantizers.57 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.13e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.down_proj_weight_quantizers.58 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.down_proj_weight_quantizers.59 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.down_proj_weight_quantizers.60 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.30e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.down_proj_weight_quantizers.61 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.99e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.down_proj_weight_quantizers.62 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.down_proj_weight_quantizers.63 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.down_proj_weight_quantizers.64 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.down_proj_weight_quantizers.65 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.down_proj_weight_quantizers.66 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.down_proj_weight_quantizers.67 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.down_proj_weight_quantizers.68 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.down_proj_weight_quantizers.69 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.65e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.down_proj_weight_quantizers.70 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.67e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.down_proj_weight_quantizers.71 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.down_proj_weight_quantizers.72 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.down_proj_weight_quantizers.73 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.down_proj_weight_quantizers.74 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.16e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.down_proj_weight_quantizers.75 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.20e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.down_proj_weight_quantizers.76 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.down_proj_weight_quantizers.77 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.down_proj_weight_quantizers.78 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.down_proj_weight_quantizers.79 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.down_proj_weight_quantizers.80 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.down_proj_weight_quantizers.81 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.down_proj_weight_quantizers.82 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.19e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.down_proj_weight_quantizers.83 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.down_proj_weight_quantizers.84 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.34e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.down_proj_weight_quantizers.85 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.26e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.down_proj_weight_quantizers.86 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.37e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.down_proj_weight_quantizers.87 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.down_proj_weight_quantizers.88 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.16e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.down_proj_weight_quantizers.89 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.down_proj_weight_quantizers.90 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.down_proj_weight_quantizers.91 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.down_proj_weight_quantizers.92 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.29e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.down_proj_weight_quantizers.93 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.down_proj_weight_quantizers.94 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.down_proj_weight_quantizers.95 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.down_proj_weight_quantizers.96 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.43e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.down_proj_weight_quantizers.97 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.down_proj_weight_quantizers.98 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.down_proj_weight_quantizers.99 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.down_proj_weight_quantizers.100 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.13e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.down_proj_weight_quantizers.101 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.down_proj_weight_quantizers.102 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.12e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.down_proj_weight_quantizers.103 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.down_proj_weight_quantizers.104 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.down_proj_weight_quantizers.105 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.down_proj_weight_quantizers.106 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.46e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.down_proj_weight_quantizers.107 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.down_proj_weight_quantizers.108 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.down_proj_weight_quantizers.109 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.16e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.down_proj_weight_quantizers.110 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.down_proj_weight_quantizers.111 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.down_proj_weight_quantizers.112 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.down_proj_weight_quantizers.113 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.down_proj_weight_quantizers.114 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.15e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.down_proj_weight_quantizers.115 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.down_proj_weight_quantizers.116 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.16e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.down_proj_weight_quantizers.117 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.down_proj_weight_quantizers.118 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.down_proj_weight_quantizers.119 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.down_proj_weight_quantizers.120 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.down_proj_weight_quantizers.121 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.down_proj_weight_quantizers.122 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.down_proj_weight_quantizers.123 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.down_proj_weight_quantizers.124 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.down_proj_weight_quantizers.125 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.down_proj_weight_quantizers.126 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.down_proj_weight_quantizers.127 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.down_proj_weight_quantizers.128 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.down_proj_weight_quantizers.129 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.29e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.down_proj_weight_quantizers.130 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.37e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.down_proj_weight_quantizers.131 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.down_proj_weight_quantizers.132 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.17e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.down_proj_weight_quantizers.133 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.down_proj_weight_quantizers.134 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.37e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.down_proj_weight_quantizers.135 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.down_proj_weight_quantizers.136 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.down_proj_weight_quantizers.137 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.26e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.down_proj_weight_quantizers.138 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.down_proj_weight_quantizers.139 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.down_proj_weight_quantizers.140 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.down_proj_weight_quantizers.141 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.66e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.down_proj_weight_quantizers.142 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.down_proj_weight_quantizers.143 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.down_proj_weight_quantizers.144 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.00e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.down_proj_weight_quantizers.145 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.down_proj_weight_quantizers.146 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.26e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.down_proj_weight_quantizers.147 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.17e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.down_proj_weight_quantizers.148 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.38e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.down_proj_weight_quantizers.149 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.84e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.down_proj_weight_quantizers.150 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.20e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.down_proj_weight_quantizers.151 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.37e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.down_proj_weight_quantizers.152 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.down_proj_weight_quantizers.153 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.down_proj_weight_quantizers.154 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.down_proj_weight_quantizers.155 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.48e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.down_proj_weight_quantizers.156 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.down_proj_weight_quantizers.157 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.down_proj_weight_quantizers.158 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.44e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.down_proj_weight_quantizers.159 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.down_proj_weight_quantizers.160 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.15e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.down_proj_weight_quantizers.161 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.67e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.down_proj_weight_quantizers.162 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.down_proj_weight_quantizers.163 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.24e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.down_proj_weight_quantizers.164 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.down_proj_weight_quantizers.165 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.down_proj_weight_quantizers.166 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.down_proj_weight_quantizers.167 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.34e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.down_proj_weight_quantizers.168 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.down_proj_weight_quantizers.169 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.down_proj_weight_quantizers.170 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.down_proj_weight_quantizers.171 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.down_proj_weight_quantizers.172 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.20e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.down_proj_weight_quantizers.173 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.down_proj_weight_quantizers.174 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.down_proj_weight_quantizers.175 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.19e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.down_proj_weight_quantizers.176 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.17e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.down_proj_weight_quantizers.177 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.down_proj_weight_quantizers.178 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.14e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.down_proj_weight_quantizers.179 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.14e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.down_proj_weight_quantizers.180 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.down_proj_weight_quantizers.181 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.down_proj_weight_quantizers.182 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.25e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.down_proj_weight_quantizers.183 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.24e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.down_proj_weight_quantizers.184 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.49e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.down_proj_weight_quantizers.185 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.down_proj_weight_quantizers.186 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.down_proj_weight_quantizers.187 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.72e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.down_proj_weight_quantizers.188 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.down_proj_weight_quantizers.189 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.13e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.down_proj_weight_quantizers.190 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.down_proj_weight_quantizers.191 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.down_proj_weight_quantizers.192 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.85e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.down_proj_weight_quantizers.193 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.down_proj_weight_quantizers.194 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.down_proj_weight_quantizers.195 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.down_proj_weight_quantizers.196 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.down_proj_weight_quantizers.197 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.down_proj_weight_quantizers.198 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.01e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.down_proj_weight_quantizers.199 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.down_proj_weight_quantizers.200 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.49e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.down_proj_weight_quantizers.201 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.down_proj_weight_quantizers.202 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.34e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.down_proj_weight_quantizers.203 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.down_proj_weight_quantizers.204 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.18e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.down_proj_weight_quantizers.205 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.down_proj_weight_quantizers.206 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.down_proj_weight_quantizers.207 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.23e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.down_proj_weight_quantizers.208 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.down_proj_weight_quantizers.209 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.down_proj_weight_quantizers.210 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.down_proj_weight_quantizers.211 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.22e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.down_proj_weight_quantizers.212 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.18e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.down_proj_weight_quantizers.213 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.00e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.down_proj_weight_quantizers.214 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.64e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.down_proj_weight_quantizers.215 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.down_proj_weight_quantizers.216 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.35e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.down_proj_weight_quantizers.217 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.down_proj_weight_quantizers.218 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.down_proj_weight_quantizers.219 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.down_proj_weight_quantizers.220 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.down_proj_weight_quantizers.221 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.down_proj_weight_quantizers.222 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.down_proj_weight_quantizers.223 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.down_proj_weight_quantizers.224 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.down_proj_weight_quantizers.225 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.35e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.down_proj_weight_quantizers.226 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.down_proj_weight_quantizers.227 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.down_proj_weight_quantizers.228 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.down_proj_weight_quantizers.229 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.19e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.down_proj_weight_quantizers.230 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.down_proj_weight_quantizers.231 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.down_proj_weight_quantizers.232 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.19e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.down_proj_weight_quantizers.233 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.down_proj_weight_quantizers.234 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.down_proj_weight_quantizers.235 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.80e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.down_proj_weight_quantizers.236 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.down_proj_weight_quantizers.237 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.down_proj_weight_quantizers.238 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.24e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.down_proj_weight_quantizers.239 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.down_proj_weight_quantizers.240 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.85e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.down_proj_weight_quantizers.241 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.25e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.down_proj_weight_quantizers.242 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.37e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.down_proj_weight_quantizers.243 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.34e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.down_proj_weight_quantizers.244 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.28e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.down_proj_weight_quantizers.245 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.down_proj_weight_quantizers.246 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.32e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.down_proj_weight_quantizers.247 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.40e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.down_proj_weight_quantizers.248 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=4.43e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.down_proj_weight_quantizers.249 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.52e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.down_proj_weight_quantizers.250 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.18e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.down_proj_weight_quantizers.251 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.38e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.down_proj_weight_quantizers.252 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.down_proj_weight_quantizers.253 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.down_proj_weight_quantizers.254 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.down_proj_weight_quantizers.255 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.shared_experts.gate_proj.input_quantizer TensorQuantizer(disabled) +model.layers.74.mlp.shared_experts.gate_proj.output_quantizer TensorQuantizer(disabled) +model.layers.74.mlp.shared_experts.gate_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.74.mlp.shared_experts.up_proj.input_quantizer TensorQuantizer(disabled) +model.layers.74.mlp.shared_experts.up_proj.output_quantizer TensorQuantizer(disabled) +model.layers.74.mlp.shared_experts.up_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.74.mlp.shared_experts.down_proj.input_quantizer TensorQuantizer(disabled) +model.layers.74.mlp.shared_experts.down_proj.output_quantizer TensorQuantizer(disabled) +model.layers.74.mlp.shared_experts.down_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.75.self_attn.q_a_proj.input_quantizer TensorQuantizer(disabled) +model.layers.75.self_attn.q_a_proj.output_quantizer TensorQuantizer(disabled) +model.layers.75.self_attn.q_a_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.75.self_attn.q_b_proj.input_quantizer TensorQuantizer(disabled) +model.layers.75.self_attn.q_b_proj.output_quantizer TensorQuantizer(disabled) +model.layers.75.self_attn.q_b_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.75.self_attn.kv_a_proj_with_mqa.input_quantizer TensorQuantizer(disabled) +model.layers.75.self_attn.kv_a_proj_with_mqa.output_quantizer TensorQuantizer(disabled) +model.layers.75.self_attn.kv_a_proj_with_mqa.weight_quantizer TensorQuantizer(disabled) +model.layers.75.self_attn.kv_b_proj.input_quantizer TensorQuantizer(disabled) +model.layers.75.self_attn.kv_b_proj.output_quantizer TensorQuantizer(disabled) +model.layers.75.self_attn.kv_b_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.75.self_attn.o_proj.input_quantizer TensorQuantizer(disabled) +model.layers.75.self_attn.o_proj.output_quantizer TensorQuantizer(disabled) +model.layers.75.self_attn.o_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.75.self_attn.q_bmm_quantizer TensorQuantizer(disabled) +model.layers.75.self_attn.k_bmm_quantizer TensorQuantizer((4, 3) bit fake per-tensor amax=4.48e+02(const) calibrator=MaxCalibrator quant) +model.layers.75.self_attn.v_bmm_quantizer TensorQuantizer((4, 3) bit fake per-tensor amax=4.48e+02(const) calibrator=MaxCalibrator quant) +model.layers.75.self_attn.p_bmm_quantizer TensorQuantizer(disabled) +model.layers.75.mlp.experts.gate_up_proj_input_quantizer TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=4.19e+00 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.gate_up_proj_weight_quantizers.0 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.25e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.gate_up_proj_weight_quantizers.1 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.25e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.gate_up_proj_weight_quantizers.2 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.17e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.gate_up_proj_weight_quantizers.3 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.51e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.gate_up_proj_weight_quantizers.4 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.gate_up_proj_weight_quantizers.5 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.23e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.gate_up_proj_weight_quantizers.6 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.gate_up_proj_weight_quantizers.7 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.11e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.gate_up_proj_weight_quantizers.8 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.gate_up_proj_weight_quantizers.9 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.36e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.gate_up_proj_weight_quantizers.10 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=4.16e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.gate_up_proj_weight_quantizers.11 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.gate_up_proj_weight_quantizers.12 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.16e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.gate_up_proj_weight_quantizers.13 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.53e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.gate_up_proj_weight_quantizers.14 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.15e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.gate_up_proj_weight_quantizers.15 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.56e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.gate_up_proj_weight_quantizers.16 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.12e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.gate_up_proj_weight_quantizers.17 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.gate_up_proj_weight_quantizers.18 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.gate_up_proj_weight_quantizers.19 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.gate_up_proj_weight_quantizers.20 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.gate_up_proj_weight_quantizers.21 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.48e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.gate_up_proj_weight_quantizers.22 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.gate_up_proj_weight_quantizers.23 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.37e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.gate_up_proj_weight_quantizers.24 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.42e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.gate_up_proj_weight_quantizers.25 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.gate_up_proj_weight_quantizers.26 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.42e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.gate_up_proj_weight_quantizers.27 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.48e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.gate_up_proj_weight_quantizers.28 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.19e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.gate_up_proj_weight_quantizers.29 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.47e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.gate_up_proj_weight_quantizers.30 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.88e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.gate_up_proj_weight_quantizers.31 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.25e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.gate_up_proj_weight_quantizers.32 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.gate_up_proj_weight_quantizers.33 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.19e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.gate_up_proj_weight_quantizers.34 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.gate_up_proj_weight_quantizers.35 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.gate_up_proj_weight_quantizers.36 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.gate_up_proj_weight_quantizers.37 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.20e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.gate_up_proj_weight_quantizers.38 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.19e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.gate_up_proj_weight_quantizers.39 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=4.22e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.gate_up_proj_weight_quantizers.40 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.gate_up_proj_weight_quantizers.41 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.46e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.gate_up_proj_weight_quantizers.42 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.gate_up_proj_weight_quantizers.43 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.13e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.gate_up_proj_weight_quantizers.44 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.30e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.gate_up_proj_weight_quantizers.45 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.38e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.gate_up_proj_weight_quantizers.46 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.gate_up_proj_weight_quantizers.47 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.38e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.gate_up_proj_weight_quantizers.48 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.26e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.gate_up_proj_weight_quantizers.49 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.24e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.gate_up_proj_weight_quantizers.50 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.68e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.gate_up_proj_weight_quantizers.51 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.15e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.gate_up_proj_weight_quantizers.52 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.35e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.gate_up_proj_weight_quantizers.53 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.gate_up_proj_weight_quantizers.54 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.gate_up_proj_weight_quantizers.55 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.gate_up_proj_weight_quantizers.56 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.28e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.gate_up_proj_weight_quantizers.57 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.23e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.gate_up_proj_weight_quantizers.58 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.98e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.gate_up_proj_weight_quantizers.59 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.48e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.gate_up_proj_weight_quantizers.60 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.17e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.gate_up_proj_weight_quantizers.61 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.gate_up_proj_weight_quantizers.62 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.17e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.gate_up_proj_weight_quantizers.63 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.21e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.gate_up_proj_weight_quantizers.64 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.54e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.gate_up_proj_weight_quantizers.65 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.gate_up_proj_weight_quantizers.66 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.17e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.gate_up_proj_weight_quantizers.67 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.54e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.gate_up_proj_weight_quantizers.68 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.20e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.gate_up_proj_weight_quantizers.69 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.gate_up_proj_weight_quantizers.70 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.36e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.gate_up_proj_weight_quantizers.71 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.27e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.gate_up_proj_weight_quantizers.72 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.60e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.gate_up_proj_weight_quantizers.73 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.25e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.gate_up_proj_weight_quantizers.74 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.gate_up_proj_weight_quantizers.75 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.gate_up_proj_weight_quantizers.76 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.29e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.gate_up_proj_weight_quantizers.77 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.18e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.gate_up_proj_weight_quantizers.78 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.gate_up_proj_weight_quantizers.79 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.18e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.gate_up_proj_weight_quantizers.80 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.28e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.gate_up_proj_weight_quantizers.81 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.15e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.gate_up_proj_weight_quantizers.82 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.45e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.gate_up_proj_weight_quantizers.83 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.20e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.gate_up_proj_weight_quantizers.84 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.26e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.gate_up_proj_weight_quantizers.85 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=4.75e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.gate_up_proj_weight_quantizers.86 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.41e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.gate_up_proj_weight_quantizers.87 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.gate_up_proj_weight_quantizers.88 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.gate_up_proj_weight_quantizers.89 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.gate_up_proj_weight_quantizers.90 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.gate_up_proj_weight_quantizers.91 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.82e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.gate_up_proj_weight_quantizers.92 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.51e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.gate_up_proj_weight_quantizers.93 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.45e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.gate_up_proj_weight_quantizers.94 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.gate_up_proj_weight_quantizers.95 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.17e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.gate_up_proj_weight_quantizers.96 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.gate_up_proj_weight_quantizers.97 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.14e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.gate_up_proj_weight_quantizers.98 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.gate_up_proj_weight_quantizers.99 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.gate_up_proj_weight_quantizers.100 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.27e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.gate_up_proj_weight_quantizers.101 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.19e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.gate_up_proj_weight_quantizers.102 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.gate_up_proj_weight_quantizers.103 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.56e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.gate_up_proj_weight_quantizers.104 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.50e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.gate_up_proj_weight_quantizers.105 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.60e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.gate_up_proj_weight_quantizers.106 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.gate_up_proj_weight_quantizers.107 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.27e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.gate_up_proj_weight_quantizers.108 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.gate_up_proj_weight_quantizers.109 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.09e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.gate_up_proj_weight_quantizers.110 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.74e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.gate_up_proj_weight_quantizers.111 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.29e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.gate_up_proj_weight_quantizers.112 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.gate_up_proj_weight_quantizers.113 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.32e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.gate_up_proj_weight_quantizers.114 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.15e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.gate_up_proj_weight_quantizers.115 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.gate_up_proj_weight_quantizers.116 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.29e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.gate_up_proj_weight_quantizers.117 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.gate_up_proj_weight_quantizers.118 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.47e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.gate_up_proj_weight_quantizers.119 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.17e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.gate_up_proj_weight_quantizers.120 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.40e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.gate_up_proj_weight_quantizers.121 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.46e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.gate_up_proj_weight_quantizers.122 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.30e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.gate_up_proj_weight_quantizers.123 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.15e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.gate_up_proj_weight_quantizers.124 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.22e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.gate_up_proj_weight_quantizers.125 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.gate_up_proj_weight_quantizers.126 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.83e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.gate_up_proj_weight_quantizers.127 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.79e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.gate_up_proj_weight_quantizers.128 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.gate_up_proj_weight_quantizers.129 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.53e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.gate_up_proj_weight_quantizers.130 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.gate_up_proj_weight_quantizers.131 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.gate_up_proj_weight_quantizers.132 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.37e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.gate_up_proj_weight_quantizers.133 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.44e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.gate_up_proj_weight_quantizers.134 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.94e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.gate_up_proj_weight_quantizers.135 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.56e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.gate_up_proj_weight_quantizers.136 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.gate_up_proj_weight_quantizers.137 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.gate_up_proj_weight_quantizers.138 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.gate_up_proj_weight_quantizers.139 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.15e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.gate_up_proj_weight_quantizers.140 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.17e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.gate_up_proj_weight_quantizers.141 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.19e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.gate_up_proj_weight_quantizers.142 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.gate_up_proj_weight_quantizers.143 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.gate_up_proj_weight_quantizers.144 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.13e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.gate_up_proj_weight_quantizers.145 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.22e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.gate_up_proj_weight_quantizers.146 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.31e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.gate_up_proj_weight_quantizers.147 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.18e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.gate_up_proj_weight_quantizers.148 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.57e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.gate_up_proj_weight_quantizers.149 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.gate_up_proj_weight_quantizers.150 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.28e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.gate_up_proj_weight_quantizers.151 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.gate_up_proj_weight_quantizers.152 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.17e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.gate_up_proj_weight_quantizers.153 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.28e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.gate_up_proj_weight_quantizers.154 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.02e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.gate_up_proj_weight_quantizers.155 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.35e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.gate_up_proj_weight_quantizers.156 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.gate_up_proj_weight_quantizers.157 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.46e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.gate_up_proj_weight_quantizers.158 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.gate_up_proj_weight_quantizers.159 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.gate_up_proj_weight_quantizers.160 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.gate_up_proj_weight_quantizers.161 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.gate_up_proj_weight_quantizers.162 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.18e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.gate_up_proj_weight_quantizers.163 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.21e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.gate_up_proj_weight_quantizers.164 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.gate_up_proj_weight_quantizers.165 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.gate_up_proj_weight_quantizers.166 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.gate_up_proj_weight_quantizers.167 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.25e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.gate_up_proj_weight_quantizers.168 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.gate_up_proj_weight_quantizers.169 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.gate_up_proj_weight_quantizers.170 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.53e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.gate_up_proj_weight_quantizers.171 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.43e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.gate_up_proj_weight_quantizers.172 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.61e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.gate_up_proj_weight_quantizers.173 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.20e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.gate_up_proj_weight_quantizers.174 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.gate_up_proj_weight_quantizers.175 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.gate_up_proj_weight_quantizers.176 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.30e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.gate_up_proj_weight_quantizers.177 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.gate_up_proj_weight_quantizers.178 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.00e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.gate_up_proj_weight_quantizers.179 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.34e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.gate_up_proj_weight_quantizers.180 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.gate_up_proj_weight_quantizers.181 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.82e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.gate_up_proj_weight_quantizers.182 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.14e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.gate_up_proj_weight_quantizers.183 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.gate_up_proj_weight_quantizers.184 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.41e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.gate_up_proj_weight_quantizers.185 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.gate_up_proj_weight_quantizers.186 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.41e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.gate_up_proj_weight_quantizers.187 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.gate_up_proj_weight_quantizers.188 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.gate_up_proj_weight_quantizers.189 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.37e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.gate_up_proj_weight_quantizers.190 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.59e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.gate_up_proj_weight_quantizers.191 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.31e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.gate_up_proj_weight_quantizers.192 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.gate_up_proj_weight_quantizers.193 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.gate_up_proj_weight_quantizers.194 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.28e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.gate_up_proj_weight_quantizers.195 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.24e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.gate_up_proj_weight_quantizers.196 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.22e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.gate_up_proj_weight_quantizers.197 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.05e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.gate_up_proj_weight_quantizers.198 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.20e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.gate_up_proj_weight_quantizers.199 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.29e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.gate_up_proj_weight_quantizers.200 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.17e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.gate_up_proj_weight_quantizers.201 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.gate_up_proj_weight_quantizers.202 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.35e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.gate_up_proj_weight_quantizers.203 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.15e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.gate_up_proj_weight_quantizers.204 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.17e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.gate_up_proj_weight_quantizers.205 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.16e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.gate_up_proj_weight_quantizers.206 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.18e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.gate_up_proj_weight_quantizers.207 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.gate_up_proj_weight_quantizers.208 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.22e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.gate_up_proj_weight_quantizers.209 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.42e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.gate_up_proj_weight_quantizers.210 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.35e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.gate_up_proj_weight_quantizers.211 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.30e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.gate_up_proj_weight_quantizers.212 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.20e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.gate_up_proj_weight_quantizers.213 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.gate_up_proj_weight_quantizers.214 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.gate_up_proj_weight_quantizers.215 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.44e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.gate_up_proj_weight_quantizers.216 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.15e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.gate_up_proj_weight_quantizers.217 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.gate_up_proj_weight_quantizers.218 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.16e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.gate_up_proj_weight_quantizers.219 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.41e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.gate_up_proj_weight_quantizers.220 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.20e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.gate_up_proj_weight_quantizers.221 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.30e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.gate_up_proj_weight_quantizers.222 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.70e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.gate_up_proj_weight_quantizers.223 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.30e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.gate_up_proj_weight_quantizers.224 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.gate_up_proj_weight_quantizers.225 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.66e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.gate_up_proj_weight_quantizers.226 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.gate_up_proj_weight_quantizers.227 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.gate_up_proj_weight_quantizers.228 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.gate_up_proj_weight_quantizers.229 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.21e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.gate_up_proj_weight_quantizers.230 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.91e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.gate_up_proj_weight_quantizers.231 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.gate_up_proj_weight_quantizers.232 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.30e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.gate_up_proj_weight_quantizers.233 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.17e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.gate_up_proj_weight_quantizers.234 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.17e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.gate_up_proj_weight_quantizers.235 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.gate_up_proj_weight_quantizers.236 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.31e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.gate_up_proj_weight_quantizers.237 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.63e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.gate_up_proj_weight_quantizers.238 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.41e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.gate_up_proj_weight_quantizers.239 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.gate_up_proj_weight_quantizers.240 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.54e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.gate_up_proj_weight_quantizers.241 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.37e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.gate_up_proj_weight_quantizers.242 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.29e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.gate_up_proj_weight_quantizers.243 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.25e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.gate_up_proj_weight_quantizers.244 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.32e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.gate_up_proj_weight_quantizers.245 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.20e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.gate_up_proj_weight_quantizers.246 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=5.43e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.gate_up_proj_weight_quantizers.247 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.gate_up_proj_weight_quantizers.248 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.gate_up_proj_weight_quantizers.249 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.gate_up_proj_weight_quantizers.250 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.gate_up_proj_weight_quantizers.251 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.gate_up_proj_weight_quantizers.252 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.25e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.gate_up_proj_weight_quantizers.253 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.gate_up_proj_weight_quantizers.254 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.gate_up_proj_weight_quantizers.255 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.down_proj_input_quantizer TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.80e+02 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.down_proj_weight_quantizers.0 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.down_proj_weight_quantizers.1 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.down_proj_weight_quantizers.2 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.82e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.down_proj_weight_quantizers.3 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.23e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.down_proj_weight_quantizers.4 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.down_proj_weight_quantizers.5 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.down_proj_weight_quantizers.6 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.down_proj_weight_quantizers.7 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.83e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.down_proj_weight_quantizers.8 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.down_proj_weight_quantizers.9 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.50e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.down_proj_weight_quantizers.10 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=4.79e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.down_proj_weight_quantizers.11 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.down_proj_weight_quantizers.12 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.down_proj_weight_quantizers.13 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.down_proj_weight_quantizers.14 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.down_proj_weight_quantizers.15 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.down_proj_weight_quantizers.16 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.46e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.down_proj_weight_quantizers.17 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.down_proj_weight_quantizers.18 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.down_proj_weight_quantizers.19 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.14e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.down_proj_weight_quantizers.20 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.down_proj_weight_quantizers.21 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.24e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.down_proj_weight_quantizers.22 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.down_proj_weight_quantizers.23 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.27e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.down_proj_weight_quantizers.24 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.13e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.down_proj_weight_quantizers.25 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.down_proj_weight_quantizers.26 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.down_proj_weight_quantizers.27 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.18e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.down_proj_weight_quantizers.28 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.16e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.down_proj_weight_quantizers.29 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.20e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.down_proj_weight_quantizers.30 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.80e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.down_proj_weight_quantizers.31 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.35e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.down_proj_weight_quantizers.32 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.down_proj_weight_quantizers.33 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.down_proj_weight_quantizers.34 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.down_proj_weight_quantizers.35 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.down_proj_weight_quantizers.36 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.13e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.down_proj_weight_quantizers.37 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.down_proj_weight_quantizers.38 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.48e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.down_proj_weight_quantizers.39 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=4.02e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.down_proj_weight_quantizers.40 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.56e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.down_proj_weight_quantizers.41 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.27e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.down_proj_weight_quantizers.42 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.down_proj_weight_quantizers.43 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.15e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.down_proj_weight_quantizers.44 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.63e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.down_proj_weight_quantizers.45 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.down_proj_weight_quantizers.46 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.down_proj_weight_quantizers.47 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=4.36e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.down_proj_weight_quantizers.48 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.30e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.down_proj_weight_quantizers.49 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.15e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.down_proj_weight_quantizers.50 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.down_proj_weight_quantizers.51 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.down_proj_weight_quantizers.52 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.down_proj_weight_quantizers.53 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.17e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.down_proj_weight_quantizers.54 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.down_proj_weight_quantizers.55 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.17e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.down_proj_weight_quantizers.56 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.down_proj_weight_quantizers.57 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.78e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.down_proj_weight_quantizers.58 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=5.12e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.down_proj_weight_quantizers.59 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.21e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.down_proj_weight_quantizers.60 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.33e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.down_proj_weight_quantizers.61 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.down_proj_weight_quantizers.62 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=4.41e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.down_proj_weight_quantizers.63 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.down_proj_weight_quantizers.64 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.15e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.down_proj_weight_quantizers.65 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.71e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.down_proj_weight_quantizers.66 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.down_proj_weight_quantizers.67 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.29e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.down_proj_weight_quantizers.68 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.44e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.down_proj_weight_quantizers.69 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.down_proj_weight_quantizers.70 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.44e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.down_proj_weight_quantizers.71 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.down_proj_weight_quantizers.72 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.53e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.down_proj_weight_quantizers.73 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.23e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.down_proj_weight_quantizers.74 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.down_proj_weight_quantizers.75 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.down_proj_weight_quantizers.76 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.down_proj_weight_quantizers.77 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.down_proj_weight_quantizers.78 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.19e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.down_proj_weight_quantizers.79 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.37e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.down_proj_weight_quantizers.80 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.25e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.down_proj_weight_quantizers.81 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.32e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.down_proj_weight_quantizers.82 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.03e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.down_proj_weight_quantizers.83 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.15e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.down_proj_weight_quantizers.84 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.28e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.down_proj_weight_quantizers.85 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.59e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.down_proj_weight_quantizers.86 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.down_proj_weight_quantizers.87 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.down_proj_weight_quantizers.88 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.21e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.down_proj_weight_quantizers.89 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.down_proj_weight_quantizers.90 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.down_proj_weight_quantizers.91 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.down_proj_weight_quantizers.92 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.68e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.down_proj_weight_quantizers.93 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.down_proj_weight_quantizers.94 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.down_proj_weight_quantizers.95 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.94e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.down_proj_weight_quantizers.96 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.down_proj_weight_quantizers.97 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.64e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.down_proj_weight_quantizers.98 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.down_proj_weight_quantizers.99 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.down_proj_weight_quantizers.100 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.19e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.down_proj_weight_quantizers.101 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.down_proj_weight_quantizers.102 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.down_proj_weight_quantizers.103 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.62e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.down_proj_weight_quantizers.104 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.22e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.down_proj_weight_quantizers.105 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.down_proj_weight_quantizers.106 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.down_proj_weight_quantizers.107 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.16e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.down_proj_weight_quantizers.108 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.down_proj_weight_quantizers.109 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.75e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.down_proj_weight_quantizers.110 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.56e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.down_proj_weight_quantizers.111 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.39e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.down_proj_weight_quantizers.112 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.down_proj_weight_quantizers.113 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.60e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.down_proj_weight_quantizers.114 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.down_proj_weight_quantizers.115 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.31e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.down_proj_weight_quantizers.116 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.down_proj_weight_quantizers.117 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.down_proj_weight_quantizers.118 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.68e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.down_proj_weight_quantizers.119 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.20e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.down_proj_weight_quantizers.120 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.17e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.down_proj_weight_quantizers.121 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.down_proj_weight_quantizers.122 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.down_proj_weight_quantizers.123 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.down_proj_weight_quantizers.124 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.55e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.down_proj_weight_quantizers.125 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.down_proj_weight_quantizers.126 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.83e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.down_proj_weight_quantizers.127 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.17e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.down_proj_weight_quantizers.128 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.down_proj_weight_quantizers.129 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.down_proj_weight_quantizers.130 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.down_proj_weight_quantizers.131 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.down_proj_weight_quantizers.132 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.17e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.down_proj_weight_quantizers.133 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.down_proj_weight_quantizers.134 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.66e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.down_proj_weight_quantizers.135 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.15e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.down_proj_weight_quantizers.136 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.20e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.down_proj_weight_quantizers.137 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.54e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.down_proj_weight_quantizers.138 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.34e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.down_proj_weight_quantizers.139 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.30e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.down_proj_weight_quantizers.140 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.down_proj_weight_quantizers.141 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.down_proj_weight_quantizers.142 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.down_proj_weight_quantizers.143 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.down_proj_weight_quantizers.144 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.down_proj_weight_quantizers.145 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.23e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.down_proj_weight_quantizers.146 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.down_proj_weight_quantizers.147 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.15e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.down_proj_weight_quantizers.148 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.down_proj_weight_quantizers.149 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.51e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.down_proj_weight_quantizers.150 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.down_proj_weight_quantizers.151 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.down_proj_weight_quantizers.152 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.13e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.down_proj_weight_quantizers.153 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.down_proj_weight_quantizers.154 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.20e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.down_proj_weight_quantizers.155 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.00e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.down_proj_weight_quantizers.156 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.down_proj_weight_quantizers.157 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.05e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.down_proj_weight_quantizers.158 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.45e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.down_proj_weight_quantizers.159 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.down_proj_weight_quantizers.160 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.down_proj_weight_quantizers.161 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.down_proj_weight_quantizers.162 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.down_proj_weight_quantizers.163 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.down_proj_weight_quantizers.164 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.down_proj_weight_quantizers.165 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.down_proj_weight_quantizers.166 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.down_proj_weight_quantizers.167 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.56e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.down_proj_weight_quantizers.168 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.down_proj_weight_quantizers.169 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.down_proj_weight_quantizers.170 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.80e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.down_proj_weight_quantizers.171 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.down_proj_weight_quantizers.172 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.60e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.down_proj_weight_quantizers.173 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.85e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.down_proj_weight_quantizers.174 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.down_proj_weight_quantizers.175 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.down_proj_weight_quantizers.176 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.down_proj_weight_quantizers.177 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.down_proj_weight_quantizers.178 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.76e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.down_proj_weight_quantizers.179 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.15e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.down_proj_weight_quantizers.180 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.48e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.down_proj_weight_quantizers.181 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.76e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.down_proj_weight_quantizers.182 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.down_proj_weight_quantizers.183 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.down_proj_weight_quantizers.184 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.down_proj_weight_quantizers.185 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.15e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.down_proj_weight_quantizers.186 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.18e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.down_proj_weight_quantizers.187 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.down_proj_weight_quantizers.188 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.down_proj_weight_quantizers.189 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.down_proj_weight_quantizers.190 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.down_proj_weight_quantizers.191 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.down_proj_weight_quantizers.192 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.down_proj_weight_quantizers.193 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.down_proj_weight_quantizers.194 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.34e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.down_proj_weight_quantizers.195 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.down_proj_weight_quantizers.196 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.17e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.down_proj_weight_quantizers.197 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.26e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.down_proj_weight_quantizers.198 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.14e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.down_proj_weight_quantizers.199 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.down_proj_weight_quantizers.200 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.down_proj_weight_quantizers.201 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.down_proj_weight_quantizers.202 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.13e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.down_proj_weight_quantizers.203 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.69e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.down_proj_weight_quantizers.204 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.down_proj_weight_quantizers.205 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.17e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.down_proj_weight_quantizers.206 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.66e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.down_proj_weight_quantizers.207 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.down_proj_weight_quantizers.208 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.14e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.down_proj_weight_quantizers.209 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.down_proj_weight_quantizers.210 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.72e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.down_proj_weight_quantizers.211 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.26e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.down_proj_weight_quantizers.212 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.20e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.down_proj_weight_quantizers.213 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.down_proj_weight_quantizers.214 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.down_proj_weight_quantizers.215 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.down_proj_weight_quantizers.216 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.down_proj_weight_quantizers.217 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.down_proj_weight_quantizers.218 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.down_proj_weight_quantizers.219 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.17e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.down_proj_weight_quantizers.220 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.14e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.down_proj_weight_quantizers.221 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.down_proj_weight_quantizers.222 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.20e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.down_proj_weight_quantizers.223 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.07e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.down_proj_weight_quantizers.224 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.down_proj_weight_quantizers.225 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.down_proj_weight_quantizers.226 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.down_proj_weight_quantizers.227 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.down_proj_weight_quantizers.228 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.down_proj_weight_quantizers.229 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.33e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.down_proj_weight_quantizers.230 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=4.32e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.down_proj_weight_quantizers.231 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.down_proj_weight_quantizers.232 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.down_proj_weight_quantizers.233 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.33e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.down_proj_weight_quantizers.234 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.down_proj_weight_quantizers.235 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.14e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.down_proj_weight_quantizers.236 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.down_proj_weight_quantizers.237 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.56e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.down_proj_weight_quantizers.238 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.down_proj_weight_quantizers.239 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.down_proj_weight_quantizers.240 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.47e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.down_proj_weight_quantizers.241 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.70e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.down_proj_weight_quantizers.242 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.13e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.down_proj_weight_quantizers.243 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.down_proj_weight_quantizers.244 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.14e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.down_proj_weight_quantizers.245 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.18e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.down_proj_weight_quantizers.246 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=4.71e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.down_proj_weight_quantizers.247 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.down_proj_weight_quantizers.248 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.down_proj_weight_quantizers.249 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.down_proj_weight_quantizers.250 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.down_proj_weight_quantizers.251 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.down_proj_weight_quantizers.252 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.down_proj_weight_quantizers.253 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.down_proj_weight_quantizers.254 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.down_proj_weight_quantizers.255 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.75.mlp.shared_experts.gate_proj.input_quantizer TensorQuantizer(disabled) +model.layers.75.mlp.shared_experts.gate_proj.output_quantizer TensorQuantizer(disabled) +model.layers.75.mlp.shared_experts.gate_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.75.mlp.shared_experts.up_proj.input_quantizer TensorQuantizer(disabled) +model.layers.75.mlp.shared_experts.up_proj.output_quantizer TensorQuantizer(disabled) +model.layers.75.mlp.shared_experts.up_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.75.mlp.shared_experts.down_proj.input_quantizer TensorQuantizer(disabled) +model.layers.75.mlp.shared_experts.down_proj.output_quantizer TensorQuantizer(disabled) +model.layers.75.mlp.shared_experts.down_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.76.self_attn.q_a_proj.input_quantizer TensorQuantizer(disabled) +model.layers.76.self_attn.q_a_proj.output_quantizer TensorQuantizer(disabled) +model.layers.76.self_attn.q_a_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.76.self_attn.q_b_proj.input_quantizer TensorQuantizer(disabled) +model.layers.76.self_attn.q_b_proj.output_quantizer TensorQuantizer(disabled) +model.layers.76.self_attn.q_b_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.76.self_attn.kv_a_proj_with_mqa.input_quantizer TensorQuantizer(disabled) +model.layers.76.self_attn.kv_a_proj_with_mqa.output_quantizer TensorQuantizer(disabled) +model.layers.76.self_attn.kv_a_proj_with_mqa.weight_quantizer TensorQuantizer(disabled) +model.layers.76.self_attn.kv_b_proj.input_quantizer TensorQuantizer(disabled) +model.layers.76.self_attn.kv_b_proj.output_quantizer TensorQuantizer(disabled) +model.layers.76.self_attn.kv_b_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.76.self_attn.o_proj.input_quantizer TensorQuantizer(disabled) +model.layers.76.self_attn.o_proj.output_quantizer TensorQuantizer(disabled) +model.layers.76.self_attn.o_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.76.self_attn.q_bmm_quantizer TensorQuantizer(disabled) +model.layers.76.self_attn.k_bmm_quantizer TensorQuantizer((4, 3) bit fake per-tensor amax=4.48e+02(const) calibrator=MaxCalibrator quant) +model.layers.76.self_attn.v_bmm_quantizer TensorQuantizer((4, 3) bit fake per-tensor amax=4.48e+02(const) calibrator=MaxCalibrator quant) +model.layers.76.self_attn.p_bmm_quantizer TensorQuantizer(disabled) +model.layers.76.mlp.experts.gate_up_proj_input_quantizer TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=7.75e+00 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.gate_up_proj_weight_quantizers.0 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.21e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.gate_up_proj_weight_quantizers.1 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.31e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.gate_up_proj_weight_quantizers.2 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.16e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.gate_up_proj_weight_quantizers.3 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.gate_up_proj_weight_quantizers.4 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.46e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.gate_up_proj_weight_quantizers.5 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.27e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.gate_up_proj_weight_quantizers.6 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.75e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.gate_up_proj_weight_quantizers.7 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.gate_up_proj_weight_quantizers.8 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.gate_up_proj_weight_quantizers.9 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.23e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.gate_up_proj_weight_quantizers.10 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.99e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.gate_up_proj_weight_quantizers.11 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.gate_up_proj_weight_quantizers.12 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.46e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.gate_up_proj_weight_quantizers.13 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.gate_up_proj_weight_quantizers.14 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.gate_up_proj_weight_quantizers.15 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.14e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.gate_up_proj_weight_quantizers.16 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.95e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.gate_up_proj_weight_quantizers.17 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.64e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.gate_up_proj_weight_quantizers.18 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.22e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.gate_up_proj_weight_quantizers.19 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.92e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.gate_up_proj_weight_quantizers.20 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.28e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.gate_up_proj_weight_quantizers.21 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.gate_up_proj_weight_quantizers.22 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.gate_up_proj_weight_quantizers.23 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.gate_up_proj_weight_quantizers.24 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.17e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.gate_up_proj_weight_quantizers.25 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.19e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.gate_up_proj_weight_quantizers.26 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.gate_up_proj_weight_quantizers.27 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.17e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.gate_up_proj_weight_quantizers.28 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.13e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.gate_up_proj_weight_quantizers.29 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.40e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.gate_up_proj_weight_quantizers.30 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.19e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.gate_up_proj_weight_quantizers.31 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.80e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.gate_up_proj_weight_quantizers.32 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.gate_up_proj_weight_quantizers.33 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.20e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.gate_up_proj_weight_quantizers.34 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.49e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.gate_up_proj_weight_quantizers.35 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.14e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.gate_up_proj_weight_quantizers.36 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.19e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.gate_up_proj_weight_quantizers.37 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.gate_up_proj_weight_quantizers.38 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.35e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.gate_up_proj_weight_quantizers.39 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.gate_up_proj_weight_quantizers.40 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.47e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.gate_up_proj_weight_quantizers.41 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.gate_up_proj_weight_quantizers.42 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.31e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.gate_up_proj_weight_quantizers.43 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.gate_up_proj_weight_quantizers.44 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.27e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.gate_up_proj_weight_quantizers.45 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.62e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.gate_up_proj_weight_quantizers.46 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.gate_up_proj_weight_quantizers.47 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.gate_up_proj_weight_quantizers.48 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.71e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.gate_up_proj_weight_quantizers.49 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.64e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.gate_up_proj_weight_quantizers.50 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.13e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.gate_up_proj_weight_quantizers.51 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.19e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.gate_up_proj_weight_quantizers.52 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.46e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.gate_up_proj_weight_quantizers.53 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.68e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.gate_up_proj_weight_quantizers.54 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.gate_up_proj_weight_quantizers.55 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.38e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.gate_up_proj_weight_quantizers.56 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.15e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.gate_up_proj_weight_quantizers.57 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.91e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.gate_up_proj_weight_quantizers.58 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.25e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.gate_up_proj_weight_quantizers.59 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.gate_up_proj_weight_quantizers.60 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.gate_up_proj_weight_quantizers.61 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.gate_up_proj_weight_quantizers.62 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.23e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.gate_up_proj_weight_quantizers.63 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.gate_up_proj_weight_quantizers.64 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.gate_up_proj_weight_quantizers.65 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.31e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.gate_up_proj_weight_quantizers.66 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.19e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.gate_up_proj_weight_quantizers.67 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.13e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.gate_up_proj_weight_quantizers.68 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.26e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.gate_up_proj_weight_quantizers.69 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.24e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.gate_up_proj_weight_quantizers.70 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.30e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.gate_up_proj_weight_quantizers.71 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.gate_up_proj_weight_quantizers.72 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.23e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.gate_up_proj_weight_quantizers.73 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.73e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.gate_up_proj_weight_quantizers.74 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.gate_up_proj_weight_quantizers.75 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.49e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.gate_up_proj_weight_quantizers.76 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.25e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.gate_up_proj_weight_quantizers.77 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.17e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.gate_up_proj_weight_quantizers.78 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.gate_up_proj_weight_quantizers.79 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.42e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.gate_up_proj_weight_quantizers.80 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.gate_up_proj_weight_quantizers.81 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.21e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.gate_up_proj_weight_quantizers.82 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.18e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.gate_up_proj_weight_quantizers.83 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.gate_up_proj_weight_quantizers.84 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.45e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.gate_up_proj_weight_quantizers.85 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.gate_up_proj_weight_quantizers.86 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=4.22e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.gate_up_proj_weight_quantizers.87 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.gate_up_proj_weight_quantizers.88 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.26e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.gate_up_proj_weight_quantizers.89 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.gate_up_proj_weight_quantizers.90 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.gate_up_proj_weight_quantizers.91 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.07e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.gate_up_proj_weight_quantizers.92 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.91e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.gate_up_proj_weight_quantizers.93 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.89e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.gate_up_proj_weight_quantizers.94 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.14e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.gate_up_proj_weight_quantizers.95 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.41e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.gate_up_proj_weight_quantizers.96 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.27e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.gate_up_proj_weight_quantizers.97 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.gate_up_proj_weight_quantizers.98 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.19e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.gate_up_proj_weight_quantizers.99 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.gate_up_proj_weight_quantizers.100 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.27e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.gate_up_proj_weight_quantizers.101 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.23e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.gate_up_proj_weight_quantizers.102 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.33e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.gate_up_proj_weight_quantizers.103 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.gate_up_proj_weight_quantizers.104 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.34e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.gate_up_proj_weight_quantizers.105 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.gate_up_proj_weight_quantizers.106 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.23e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.gate_up_proj_weight_quantizers.107 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.gate_up_proj_weight_quantizers.108 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.43e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.gate_up_proj_weight_quantizers.109 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.95e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.gate_up_proj_weight_quantizers.110 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.23e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.gate_up_proj_weight_quantizers.111 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.19e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.gate_up_proj_weight_quantizers.112 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.43e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.gate_up_proj_weight_quantizers.113 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.gate_up_proj_weight_quantizers.114 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.29e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.gate_up_proj_weight_quantizers.115 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.22e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.gate_up_proj_weight_quantizers.116 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.18e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.gate_up_proj_weight_quantizers.117 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.62e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.gate_up_proj_weight_quantizers.118 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.gate_up_proj_weight_quantizers.119 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.18e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.gate_up_proj_weight_quantizers.120 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.17e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.gate_up_proj_weight_quantizers.121 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.79e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.gate_up_proj_weight_quantizers.122 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.gate_up_proj_weight_quantizers.123 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.58e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.gate_up_proj_weight_quantizers.124 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.gate_up_proj_weight_quantizers.125 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.gate_up_proj_weight_quantizers.126 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.25e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.gate_up_proj_weight_quantizers.127 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.gate_up_proj_weight_quantizers.128 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.39e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.gate_up_proj_weight_quantizers.129 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.gate_up_proj_weight_quantizers.130 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.21e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.gate_up_proj_weight_quantizers.131 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.42e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.gate_up_proj_weight_quantizers.132 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.11e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.gate_up_proj_weight_quantizers.133 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.32e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.gate_up_proj_weight_quantizers.134 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.42e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.gate_up_proj_weight_quantizers.135 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=5.39e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.gate_up_proj_weight_quantizers.136 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.gate_up_proj_weight_quantizers.137 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.25e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.gate_up_proj_weight_quantizers.138 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.gate_up_proj_weight_quantizers.139 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.54e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.gate_up_proj_weight_quantizers.140 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.33e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.gate_up_proj_weight_quantizers.141 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.31e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.gate_up_proj_weight_quantizers.142 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.gate_up_proj_weight_quantizers.143 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.58e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.gate_up_proj_weight_quantizers.144 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.26e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.gate_up_proj_weight_quantizers.145 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.gate_up_proj_weight_quantizers.146 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.59e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.gate_up_proj_weight_quantizers.147 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.32e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.gate_up_proj_weight_quantizers.148 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.gate_up_proj_weight_quantizers.149 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.gate_up_proj_weight_quantizers.150 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.gate_up_proj_weight_quantizers.151 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.23e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.gate_up_proj_weight_quantizers.152 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.26e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.gate_up_proj_weight_quantizers.153 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.gate_up_proj_weight_quantizers.154 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.25e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.gate_up_proj_weight_quantizers.155 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.09e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.gate_up_proj_weight_quantizers.156 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.19e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.gate_up_proj_weight_quantizers.157 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.16e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.gate_up_proj_weight_quantizers.158 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.12e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.gate_up_proj_weight_quantizers.159 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.33e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.gate_up_proj_weight_quantizers.160 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.33e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.gate_up_proj_weight_quantizers.161 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.21e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.gate_up_proj_weight_quantizers.162 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.gate_up_proj_weight_quantizers.163 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.gate_up_proj_weight_quantizers.164 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.gate_up_proj_weight_quantizers.165 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.23e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.gate_up_proj_weight_quantizers.166 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.20e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.gate_up_proj_weight_quantizers.167 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.75e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.gate_up_proj_weight_quantizers.168 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.gate_up_proj_weight_quantizers.169 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.gate_up_proj_weight_quantizers.170 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.73e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.gate_up_proj_weight_quantizers.171 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.gate_up_proj_weight_quantizers.172 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.37e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.gate_up_proj_weight_quantizers.173 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.38e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.gate_up_proj_weight_quantizers.174 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.39e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.gate_up_proj_weight_quantizers.175 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.49e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.gate_up_proj_weight_quantizers.176 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.gate_up_proj_weight_quantizers.177 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.11e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.gate_up_proj_weight_quantizers.178 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.29e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.gate_up_proj_weight_quantizers.179 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.gate_up_proj_weight_quantizers.180 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.21e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.gate_up_proj_weight_quantizers.181 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.15e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.gate_up_proj_weight_quantizers.182 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.17e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.gate_up_proj_weight_quantizers.183 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.gate_up_proj_weight_quantizers.184 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.18e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.gate_up_proj_weight_quantizers.185 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.gate_up_proj_weight_quantizers.186 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.30e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.gate_up_proj_weight_quantizers.187 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.gate_up_proj_weight_quantizers.188 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.gate_up_proj_weight_quantizers.189 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.69e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.gate_up_proj_weight_quantizers.190 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.19e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.gate_up_proj_weight_quantizers.191 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.gate_up_proj_weight_quantizers.192 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.27e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.gate_up_proj_weight_quantizers.193 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.14e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.gate_up_proj_weight_quantizers.194 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.49e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.gate_up_proj_weight_quantizers.195 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.24e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.gate_up_proj_weight_quantizers.196 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.42e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.gate_up_proj_weight_quantizers.197 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.13e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.gate_up_proj_weight_quantizers.198 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.16e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.gate_up_proj_weight_quantizers.199 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.50e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.gate_up_proj_weight_quantizers.200 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.75e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.gate_up_proj_weight_quantizers.201 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.60e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.gate_up_proj_weight_quantizers.202 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.16e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.gate_up_proj_weight_quantizers.203 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.gate_up_proj_weight_quantizers.204 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.21e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.gate_up_proj_weight_quantizers.205 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.19e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.gate_up_proj_weight_quantizers.206 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.23e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.gate_up_proj_weight_quantizers.207 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.17e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.gate_up_proj_weight_quantizers.208 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.15e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.gate_up_proj_weight_quantizers.209 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.gate_up_proj_weight_quantizers.210 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.13e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.gate_up_proj_weight_quantizers.211 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.73e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.gate_up_proj_weight_quantizers.212 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.49e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.gate_up_proj_weight_quantizers.213 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.26e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.gate_up_proj_weight_quantizers.214 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.34e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.gate_up_proj_weight_quantizers.215 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=4.86e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.gate_up_proj_weight_quantizers.216 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.15e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.gate_up_proj_weight_quantizers.217 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.gate_up_proj_weight_quantizers.218 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.63e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.gate_up_proj_weight_quantizers.219 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.16e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.gate_up_proj_weight_quantizers.220 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.17e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.gate_up_proj_weight_quantizers.221 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.38e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.gate_up_proj_weight_quantizers.222 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.13e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.gate_up_proj_weight_quantizers.223 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.gate_up_proj_weight_quantizers.224 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.gate_up_proj_weight_quantizers.225 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.gate_up_proj_weight_quantizers.226 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.32e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.gate_up_proj_weight_quantizers.227 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.34e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.gate_up_proj_weight_quantizers.228 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.gate_up_proj_weight_quantizers.229 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.gate_up_proj_weight_quantizers.230 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.19e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.gate_up_proj_weight_quantizers.231 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.80e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.gate_up_proj_weight_quantizers.232 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.gate_up_proj_weight_quantizers.233 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.22e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.gate_up_proj_weight_quantizers.234 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.13e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.gate_up_proj_weight_quantizers.235 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.gate_up_proj_weight_quantizers.236 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.34e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.gate_up_proj_weight_quantizers.237 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.46e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.gate_up_proj_weight_quantizers.238 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=5.04e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.gate_up_proj_weight_quantizers.239 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.gate_up_proj_weight_quantizers.240 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.35e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.gate_up_proj_weight_quantizers.241 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.13e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.gate_up_proj_weight_quantizers.242 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.gate_up_proj_weight_quantizers.243 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.20e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.gate_up_proj_weight_quantizers.244 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.29e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.gate_up_proj_weight_quantizers.245 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.gate_up_proj_weight_quantizers.246 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.gate_up_proj_weight_quantizers.247 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.22e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.gate_up_proj_weight_quantizers.248 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.24e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.gate_up_proj_weight_quantizers.249 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.37e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.gate_up_proj_weight_quantizers.250 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.24e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.gate_up_proj_weight_quantizers.251 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.gate_up_proj_weight_quantizers.252 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.22e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.gate_up_proj_weight_quantizers.253 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.gate_up_proj_weight_quantizers.254 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.13e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.gate_up_proj_weight_quantizers.255 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.13e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.down_proj_input_quantizer TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.26e+02 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.down_proj_weight_quantizers.0 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.07e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.down_proj_weight_quantizers.1 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.down_proj_weight_quantizers.2 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.45e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.down_proj_weight_quantizers.3 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.18e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.down_proj_weight_quantizers.4 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.35e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.down_proj_weight_quantizers.5 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.down_proj_weight_quantizers.6 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.21e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.down_proj_weight_quantizers.7 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.31e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.down_proj_weight_quantizers.8 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.down_proj_weight_quantizers.9 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.15e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.down_proj_weight_quantizers.10 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.33e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.down_proj_weight_quantizers.11 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.down_proj_weight_quantizers.12 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.down_proj_weight_quantizers.13 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.down_proj_weight_quantizers.14 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.25e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.down_proj_weight_quantizers.15 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.48e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.down_proj_weight_quantizers.16 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.26e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.down_proj_weight_quantizers.17 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.54e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.down_proj_weight_quantizers.18 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.down_proj_weight_quantizers.19 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.98e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.down_proj_weight_quantizers.20 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.48e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.down_proj_weight_quantizers.21 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.54e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.down_proj_weight_quantizers.22 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.36e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.down_proj_weight_quantizers.23 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.25e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.down_proj_weight_quantizers.24 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.down_proj_weight_quantizers.25 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.15e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.down_proj_weight_quantizers.26 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.21e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.down_proj_weight_quantizers.27 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.down_proj_weight_quantizers.28 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.78e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.down_proj_weight_quantizers.29 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.63e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.down_proj_weight_quantizers.30 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.30e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.down_proj_weight_quantizers.31 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.40e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.down_proj_weight_quantizers.32 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.down_proj_weight_quantizers.33 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.17e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.down_proj_weight_quantizers.34 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.44e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.down_proj_weight_quantizers.35 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.down_proj_weight_quantizers.36 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.down_proj_weight_quantizers.37 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.down_proj_weight_quantizers.38 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.36e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.down_proj_weight_quantizers.39 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.54e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.down_proj_weight_quantizers.40 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.48e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.down_proj_weight_quantizers.41 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.down_proj_weight_quantizers.42 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.13e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.down_proj_weight_quantizers.43 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.27e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.down_proj_weight_quantizers.44 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.90e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.down_proj_weight_quantizers.45 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.down_proj_weight_quantizers.46 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.78e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.down_proj_weight_quantizers.47 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.41e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.down_proj_weight_quantizers.48 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.66e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.down_proj_weight_quantizers.49 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.40e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.down_proj_weight_quantizers.50 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.down_proj_weight_quantizers.51 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.13e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.down_proj_weight_quantizers.52 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.91e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.down_proj_weight_quantizers.53 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.30e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.down_proj_weight_quantizers.54 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.down_proj_weight_quantizers.55 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.18e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.down_proj_weight_quantizers.56 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.94e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.down_proj_weight_quantizers.57 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.91e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.down_proj_weight_quantizers.58 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.61e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.down_proj_weight_quantizers.59 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.95e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.down_proj_weight_quantizers.60 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.down_proj_weight_quantizers.61 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.21e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.down_proj_weight_quantizers.62 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.71e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.down_proj_weight_quantizers.63 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.down_proj_weight_quantizers.64 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.down_proj_weight_quantizers.65 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.down_proj_weight_quantizers.66 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.85e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.down_proj_weight_quantizers.67 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.47e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.down_proj_weight_quantizers.68 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.down_proj_weight_quantizers.69 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.28e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.down_proj_weight_quantizers.70 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.67e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.down_proj_weight_quantizers.71 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.17e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.down_proj_weight_quantizers.72 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.down_proj_weight_quantizers.73 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=4.22e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.down_proj_weight_quantizers.74 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.down_proj_weight_quantizers.75 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.19e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.down_proj_weight_quantizers.76 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.13e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.down_proj_weight_quantizers.77 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.17e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.down_proj_weight_quantizers.78 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.35e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.down_proj_weight_quantizers.79 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.down_proj_weight_quantizers.80 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.down_proj_weight_quantizers.81 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.15e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.down_proj_weight_quantizers.82 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.68e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.down_proj_weight_quantizers.83 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.26e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.down_proj_weight_quantizers.84 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.down_proj_weight_quantizers.85 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.21e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.down_proj_weight_quantizers.86 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.22e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.down_proj_weight_quantizers.87 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.24e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.down_proj_weight_quantizers.88 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.74e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.down_proj_weight_quantizers.89 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.24e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.down_proj_weight_quantizers.90 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.56e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.down_proj_weight_quantizers.91 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.29e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.down_proj_weight_quantizers.92 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.88e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.down_proj_weight_quantizers.93 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.56e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.down_proj_weight_quantizers.94 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.down_proj_weight_quantizers.95 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.96e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.down_proj_weight_quantizers.96 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.15e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.down_proj_weight_quantizers.97 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.down_proj_weight_quantizers.98 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.19e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.down_proj_weight_quantizers.99 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.14e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.down_proj_weight_quantizers.100 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.down_proj_weight_quantizers.101 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.26e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.down_proj_weight_quantizers.102 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.down_proj_weight_quantizers.103 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.down_proj_weight_quantizers.104 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.46e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.down_proj_weight_quantizers.105 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.17e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.down_proj_weight_quantizers.106 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.down_proj_weight_quantizers.107 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.16e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.down_proj_weight_quantizers.108 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.71e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.down_proj_weight_quantizers.109 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.02e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.down_proj_weight_quantizers.110 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.down_proj_weight_quantizers.111 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.down_proj_weight_quantizers.112 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.29e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.down_proj_weight_quantizers.113 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.14e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.down_proj_weight_quantizers.114 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.66e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.down_proj_weight_quantizers.115 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.85e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.down_proj_weight_quantizers.116 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.14e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.down_proj_weight_quantizers.117 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.94e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.down_proj_weight_quantizers.118 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.down_proj_weight_quantizers.119 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.down_proj_weight_quantizers.120 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.down_proj_weight_quantizers.121 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.22e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.down_proj_weight_quantizers.122 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.33e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.down_proj_weight_quantizers.123 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.down_proj_weight_quantizers.124 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.18e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.down_proj_weight_quantizers.125 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.down_proj_weight_quantizers.126 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.35e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.down_proj_weight_quantizers.127 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.down_proj_weight_quantizers.128 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.86e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.down_proj_weight_quantizers.129 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.down_proj_weight_quantizers.130 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.18e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.down_proj_weight_quantizers.131 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.46e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.down_proj_weight_quantizers.132 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.76e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.down_proj_weight_quantizers.133 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.23e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.down_proj_weight_quantizers.134 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.93e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.down_proj_weight_quantizers.135 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.26e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.down_proj_weight_quantizers.136 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.down_proj_weight_quantizers.137 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.22e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.down_proj_weight_quantizers.138 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.25e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.down_proj_weight_quantizers.139 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.52e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.down_proj_weight_quantizers.140 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=4.16e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.down_proj_weight_quantizers.141 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.54e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.down_proj_weight_quantizers.142 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.29e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.down_proj_weight_quantizers.143 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.43e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.down_proj_weight_quantizers.144 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.21e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.down_proj_weight_quantizers.145 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.30e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.down_proj_weight_quantizers.146 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.20e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.down_proj_weight_quantizers.147 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.22e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.down_proj_weight_quantizers.148 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.15e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.down_proj_weight_quantizers.149 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.20e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.down_proj_weight_quantizers.150 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.down_proj_weight_quantizers.151 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.17e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.down_proj_weight_quantizers.152 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.33e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.down_proj_weight_quantizers.153 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.down_proj_weight_quantizers.154 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.23e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.down_proj_weight_quantizers.155 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.57e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.down_proj_weight_quantizers.156 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.down_proj_weight_quantizers.157 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.down_proj_weight_quantizers.158 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.down_proj_weight_quantizers.159 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.23e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.down_proj_weight_quantizers.160 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.18e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.down_proj_weight_quantizers.161 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.25e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.down_proj_weight_quantizers.162 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.20e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.down_proj_weight_quantizers.163 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.down_proj_weight_quantizers.164 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.18e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.down_proj_weight_quantizers.165 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.64e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.down_proj_weight_quantizers.166 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.17e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.down_proj_weight_quantizers.167 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.88e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.down_proj_weight_quantizers.168 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.down_proj_weight_quantizers.169 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.13e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.down_proj_weight_quantizers.170 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.58e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.down_proj_weight_quantizers.171 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.16e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.down_proj_weight_quantizers.172 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.25e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.down_proj_weight_quantizers.173 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.47e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.down_proj_weight_quantizers.174 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.56e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.down_proj_weight_quantizers.175 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.37e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.down_proj_weight_quantizers.176 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.19e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.down_proj_weight_quantizers.177 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.26e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.down_proj_weight_quantizers.178 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.down_proj_weight_quantizers.179 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.down_proj_weight_quantizers.180 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.24e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.down_proj_weight_quantizers.181 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.down_proj_weight_quantizers.182 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.down_proj_weight_quantizers.183 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.down_proj_weight_quantizers.184 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.72e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.down_proj_weight_quantizers.185 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.down_proj_weight_quantizers.186 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.down_proj_weight_quantizers.187 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.down_proj_weight_quantizers.188 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.down_proj_weight_quantizers.189 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.down_proj_weight_quantizers.190 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.19e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.down_proj_weight_quantizers.191 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.down_proj_weight_quantizers.192 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.28e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.down_proj_weight_quantizers.193 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.down_proj_weight_quantizers.194 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.25e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.down_proj_weight_quantizers.195 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.23e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.down_proj_weight_quantizers.196 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.43e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.down_proj_weight_quantizers.197 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.23e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.down_proj_weight_quantizers.198 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.down_proj_weight_quantizers.199 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.29e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.down_proj_weight_quantizers.200 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.23e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.down_proj_weight_quantizers.201 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.35e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.down_proj_weight_quantizers.202 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.35e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.down_proj_weight_quantizers.203 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.16e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.down_proj_weight_quantizers.204 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.down_proj_weight_quantizers.205 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.25e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.down_proj_weight_quantizers.206 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.25e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.down_proj_weight_quantizers.207 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.55e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.down_proj_weight_quantizers.208 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.down_proj_weight_quantizers.209 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.down_proj_weight_quantizers.210 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.16e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.down_proj_weight_quantizers.211 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.71e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.down_proj_weight_quantizers.212 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.down_proj_weight_quantizers.213 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.45e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.down_proj_weight_quantizers.214 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.20e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.down_proj_weight_quantizers.215 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.05e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.down_proj_weight_quantizers.216 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.34e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.down_proj_weight_quantizers.217 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.down_proj_weight_quantizers.218 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.22e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.down_proj_weight_quantizers.219 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.down_proj_weight_quantizers.220 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.24e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.down_proj_weight_quantizers.221 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.50e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.down_proj_weight_quantizers.222 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.15e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.down_proj_weight_quantizers.223 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.down_proj_weight_quantizers.224 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.down_proj_weight_quantizers.225 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.28e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.down_proj_weight_quantizers.226 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.down_proj_weight_quantizers.227 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.down_proj_weight_quantizers.228 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.28e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.down_proj_weight_quantizers.229 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.down_proj_weight_quantizers.230 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.down_proj_weight_quantizers.231 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.down_proj_weight_quantizers.232 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.down_proj_weight_quantizers.233 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.14e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.down_proj_weight_quantizers.234 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.down_proj_weight_quantizers.235 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.17e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.down_proj_weight_quantizers.236 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=5.51e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.down_proj_weight_quantizers.237 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.33e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.down_proj_weight_quantizers.238 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.09e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.down_proj_weight_quantizers.239 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.down_proj_weight_quantizers.240 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.down_proj_weight_quantizers.241 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.35e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.down_proj_weight_quantizers.242 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.47e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.down_proj_weight_quantizers.243 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.36e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.down_proj_weight_quantizers.244 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.29e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.down_proj_weight_quantizers.245 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.down_proj_weight_quantizers.246 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.39e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.down_proj_weight_quantizers.247 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.down_proj_weight_quantizers.248 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.20e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.down_proj_weight_quantizers.249 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.30e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.down_proj_weight_quantizers.250 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.down_proj_weight_quantizers.251 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.down_proj_weight_quantizers.252 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.44e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.down_proj_weight_quantizers.253 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.down_proj_weight_quantizers.254 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.down_proj_weight_quantizers.255 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.17e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.shared_experts.gate_proj.input_quantizer TensorQuantizer(disabled) +model.layers.76.mlp.shared_experts.gate_proj.output_quantizer TensorQuantizer(disabled) +model.layers.76.mlp.shared_experts.gate_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.76.mlp.shared_experts.up_proj.input_quantizer TensorQuantizer(disabled) +model.layers.76.mlp.shared_experts.up_proj.output_quantizer TensorQuantizer(disabled) +model.layers.76.mlp.shared_experts.up_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.76.mlp.shared_experts.down_proj.input_quantizer TensorQuantizer(disabled) +model.layers.76.mlp.shared_experts.down_proj.output_quantizer TensorQuantizer(disabled) +model.layers.76.mlp.shared_experts.down_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.77.self_attn.q_a_proj.input_quantizer TensorQuantizer(disabled) +model.layers.77.self_attn.q_a_proj.output_quantizer TensorQuantizer(disabled) +model.layers.77.self_attn.q_a_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.77.self_attn.q_b_proj.input_quantizer TensorQuantizer(disabled) +model.layers.77.self_attn.q_b_proj.output_quantizer TensorQuantizer(disabled) +model.layers.77.self_attn.q_b_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.77.self_attn.kv_a_proj_with_mqa.input_quantizer TensorQuantizer(disabled) +model.layers.77.self_attn.kv_a_proj_with_mqa.output_quantizer TensorQuantizer(disabled) +model.layers.77.self_attn.kv_a_proj_with_mqa.weight_quantizer TensorQuantizer(disabled) +model.layers.77.self_attn.kv_b_proj.input_quantizer TensorQuantizer(disabled) +model.layers.77.self_attn.kv_b_proj.output_quantizer TensorQuantizer(disabled) +model.layers.77.self_attn.kv_b_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.77.self_attn.o_proj.input_quantizer TensorQuantizer(disabled) +model.layers.77.self_attn.o_proj.output_quantizer TensorQuantizer(disabled) +model.layers.77.self_attn.o_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.77.self_attn.q_bmm_quantizer TensorQuantizer(disabled) +model.layers.77.self_attn.k_bmm_quantizer TensorQuantizer((4, 3) bit fake per-tensor amax=4.48e+02(const) calibrator=MaxCalibrator quant) +model.layers.77.self_attn.v_bmm_quantizer TensorQuantizer((4, 3) bit fake per-tensor amax=4.48e+02(const) calibrator=MaxCalibrator quant) +model.layers.77.self_attn.p_bmm_quantizer TensorQuantizer(disabled) +model.layers.77.mlp.experts.gate_up_proj_input_quantizer TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=6.91e+00 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.gate_up_proj_weight_quantizers.0 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.45e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.gate_up_proj_weight_quantizers.1 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.63e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.gate_up_proj_weight_quantizers.2 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.54e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.gate_up_proj_weight_quantizers.3 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.14e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.gate_up_proj_weight_quantizers.4 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.53e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.gate_up_proj_weight_quantizers.5 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.46e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.gate_up_proj_weight_quantizers.6 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.14e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.gate_up_proj_weight_quantizers.7 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.88e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.gate_up_proj_weight_quantizers.8 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.65e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.gate_up_proj_weight_quantizers.9 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.35e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.gate_up_proj_weight_quantizers.10 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.56e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.gate_up_proj_weight_quantizers.11 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.61e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.gate_up_proj_weight_quantizers.12 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.01e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.gate_up_proj_weight_quantizers.13 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.14e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.gate_up_proj_weight_quantizers.14 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.17e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.gate_up_proj_weight_quantizers.15 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.66e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.gate_up_proj_weight_quantizers.16 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.88e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.gate_up_proj_weight_quantizers.17 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.28e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.gate_up_proj_weight_quantizers.18 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.77e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.gate_up_proj_weight_quantizers.19 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.25e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.gate_up_proj_weight_quantizers.20 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.36e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.gate_up_proj_weight_quantizers.21 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.84e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.gate_up_proj_weight_quantizers.22 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=4.10e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.gate_up_proj_weight_quantizers.23 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.19e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.gate_up_proj_weight_quantizers.24 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.77e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.gate_up_proj_weight_quantizers.25 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.46e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.gate_up_proj_weight_quantizers.26 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.85e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.gate_up_proj_weight_quantizers.27 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.36e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.gate_up_proj_weight_quantizers.28 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.87e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.gate_up_proj_weight_quantizers.29 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.35e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.gate_up_proj_weight_quantizers.30 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.20e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.gate_up_proj_weight_quantizers.31 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.51e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.gate_up_proj_weight_quantizers.32 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.34e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.gate_up_proj_weight_quantizers.33 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.92e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.gate_up_proj_weight_quantizers.34 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.29e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.gate_up_proj_weight_quantizers.35 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.25e+00 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.gate_up_proj_weight_quantizers.36 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.79e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.gate_up_proj_weight_quantizers.37 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.19e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.gate_up_proj_weight_quantizers.38 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.21e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.gate_up_proj_weight_quantizers.39 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.77e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.gate_up_proj_weight_quantizers.40 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.gate_up_proj_weight_quantizers.41 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.38e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.gate_up_proj_weight_quantizers.42 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.20e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.gate_up_proj_weight_quantizers.43 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.57e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.gate_up_proj_weight_quantizers.44 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.66e+00 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.gate_up_proj_weight_quantizers.45 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.42e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.gate_up_proj_weight_quantizers.46 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.81e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.gate_up_proj_weight_quantizers.47 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.99e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.gate_up_proj_weight_quantizers.48 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.16e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.gate_up_proj_weight_quantizers.49 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.86e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.gate_up_proj_weight_quantizers.50 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.28e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.gate_up_proj_weight_quantizers.51 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.24e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.gate_up_proj_weight_quantizers.52 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.30e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.gate_up_proj_weight_quantizers.53 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.75e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.gate_up_proj_weight_quantizers.54 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.58e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.gate_up_proj_weight_quantizers.55 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.63e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.gate_up_proj_weight_quantizers.56 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.14e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.gate_up_proj_weight_quantizers.57 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.59e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.gate_up_proj_weight_quantizers.58 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.23e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.gate_up_proj_weight_quantizers.59 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.45e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.gate_up_proj_weight_quantizers.60 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.22e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.gate_up_proj_weight_quantizers.61 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=4.28e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.gate_up_proj_weight_quantizers.62 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.gate_up_proj_weight_quantizers.63 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.46e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.gate_up_proj_weight_quantizers.64 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.80e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.gate_up_proj_weight_quantizers.65 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.49e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.gate_up_proj_weight_quantizers.66 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.31e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.gate_up_proj_weight_quantizers.67 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.87e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.gate_up_proj_weight_quantizers.68 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.58e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.gate_up_proj_weight_quantizers.69 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.64e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.gate_up_proj_weight_quantizers.70 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.29e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.gate_up_proj_weight_quantizers.71 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.34e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.gate_up_proj_weight_quantizers.72 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.37e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.gate_up_proj_weight_quantizers.73 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.70e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.gate_up_proj_weight_quantizers.74 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.49e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.gate_up_proj_weight_quantizers.75 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.33e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.gate_up_proj_weight_quantizers.76 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.86e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.gate_up_proj_weight_quantizers.77 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.gate_up_proj_weight_quantizers.78 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.71e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.gate_up_proj_weight_quantizers.79 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.67e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.gate_up_proj_weight_quantizers.80 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.59e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.gate_up_proj_weight_quantizers.81 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.52e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.gate_up_proj_weight_quantizers.82 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.64e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.gate_up_proj_weight_quantizers.83 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.56e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.gate_up_proj_weight_quantizers.84 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.62e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.gate_up_proj_weight_quantizers.85 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.21e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.gate_up_proj_weight_quantizers.86 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.70e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.gate_up_proj_weight_quantizers.87 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.38e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.gate_up_proj_weight_quantizers.88 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.64e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.gate_up_proj_weight_quantizers.89 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.49e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.gate_up_proj_weight_quantizers.90 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=4.18e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.gate_up_proj_weight_quantizers.91 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.27e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.gate_up_proj_weight_quantizers.92 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.46e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.gate_up_proj_weight_quantizers.93 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.65e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.gate_up_proj_weight_quantizers.94 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.47e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.gate_up_proj_weight_quantizers.95 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.59e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.gate_up_proj_weight_quantizers.96 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.76e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.gate_up_proj_weight_quantizers.97 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.28e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.gate_up_proj_weight_quantizers.98 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.76e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.gate_up_proj_weight_quantizers.99 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.27e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.gate_up_proj_weight_quantizers.100 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.66e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.gate_up_proj_weight_quantizers.101 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.36e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.gate_up_proj_weight_quantizers.102 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.33e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.gate_up_proj_weight_quantizers.103 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.13e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.gate_up_proj_weight_quantizers.104 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.43e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.gate_up_proj_weight_quantizers.105 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.56e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.gate_up_proj_weight_quantizers.106 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=6.37e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.gate_up_proj_weight_quantizers.107 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.68e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.gate_up_proj_weight_quantizers.108 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.33e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.gate_up_proj_weight_quantizers.109 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.82e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.gate_up_proj_weight_quantizers.110 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.63e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.gate_up_proj_weight_quantizers.111 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.40e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.gate_up_proj_weight_quantizers.112 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.19e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.gate_up_proj_weight_quantizers.113 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.11e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.gate_up_proj_weight_quantizers.114 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.71e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.gate_up_proj_weight_quantizers.115 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.37e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.gate_up_proj_weight_quantizers.116 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.73e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.gate_up_proj_weight_quantizers.117 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.42e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.gate_up_proj_weight_quantizers.118 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.82e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.gate_up_proj_weight_quantizers.119 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.81e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.gate_up_proj_weight_quantizers.120 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.31e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.gate_up_proj_weight_quantizers.121 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.46e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.gate_up_proj_weight_quantizers.122 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.09e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.gate_up_proj_weight_quantizers.123 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.45e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.gate_up_proj_weight_quantizers.124 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.35e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.gate_up_proj_weight_quantizers.125 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.90e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.gate_up_proj_weight_quantizers.126 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.69e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.gate_up_proj_weight_quantizers.127 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.50e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.gate_up_proj_weight_quantizers.128 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.gate_up_proj_weight_quantizers.129 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.25e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.gate_up_proj_weight_quantizers.130 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.47e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.gate_up_proj_weight_quantizers.131 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.33e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.gate_up_proj_weight_quantizers.132 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.20e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.gate_up_proj_weight_quantizers.133 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.87e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.gate_up_proj_weight_quantizers.134 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.11e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.gate_up_proj_weight_quantizers.135 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.29e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.gate_up_proj_weight_quantizers.136 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.32e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.gate_up_proj_weight_quantizers.137 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.65e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.gate_up_proj_weight_quantizers.138 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.19e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.gate_up_proj_weight_quantizers.139 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=4.34e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.gate_up_proj_weight_quantizers.140 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.29e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.gate_up_proj_weight_quantizers.141 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.29e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.gate_up_proj_weight_quantizers.142 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.54e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.gate_up_proj_weight_quantizers.143 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.43e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.gate_up_proj_weight_quantizers.144 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.40e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.gate_up_proj_weight_quantizers.145 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.77e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.gate_up_proj_weight_quantizers.146 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.gate_up_proj_weight_quantizers.147 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.71e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.gate_up_proj_weight_quantizers.148 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.79e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.gate_up_proj_weight_quantizers.149 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.25e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.gate_up_proj_weight_quantizers.150 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.36e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.gate_up_proj_weight_quantizers.151 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.27e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.gate_up_proj_weight_quantizers.152 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.14e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.gate_up_proj_weight_quantizers.153 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.14e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.gate_up_proj_weight_quantizers.154 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.27e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.gate_up_proj_weight_quantizers.155 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.22e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.gate_up_proj_weight_quantizers.156 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.42e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.gate_up_proj_weight_quantizers.157 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.88e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.gate_up_proj_weight_quantizers.158 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.25e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.gate_up_proj_weight_quantizers.159 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.14e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.gate_up_proj_weight_quantizers.160 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=4.16e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.gate_up_proj_weight_quantizers.161 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.55e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.gate_up_proj_weight_quantizers.162 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.31e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.gate_up_proj_weight_quantizers.163 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.91e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.gate_up_proj_weight_quantizers.164 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.40e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.gate_up_proj_weight_quantizers.165 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.25e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.gate_up_proj_weight_quantizers.166 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.91e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.gate_up_proj_weight_quantizers.167 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.50e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.gate_up_proj_weight_quantizers.168 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.43e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.gate_up_proj_weight_quantizers.169 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.gate_up_proj_weight_quantizers.170 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.47e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.gate_up_proj_weight_quantizers.171 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.40e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.gate_up_proj_weight_quantizers.172 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.46e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.gate_up_proj_weight_quantizers.173 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.66e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.gate_up_proj_weight_quantizers.174 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.18e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.gate_up_proj_weight_quantizers.175 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.19e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.gate_up_proj_weight_quantizers.176 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.18e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.gate_up_proj_weight_quantizers.177 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.82e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.gate_up_proj_weight_quantizers.178 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.49e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.gate_up_proj_weight_quantizers.179 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.16e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.gate_up_proj_weight_quantizers.180 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.gate_up_proj_weight_quantizers.181 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.20e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.gate_up_proj_weight_quantizers.182 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.28e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.gate_up_proj_weight_quantizers.183 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.51e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.gate_up_proj_weight_quantizers.184 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.20e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.gate_up_proj_weight_quantizers.185 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.78e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.gate_up_proj_weight_quantizers.186 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.16e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.gate_up_proj_weight_quantizers.187 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.64e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.gate_up_proj_weight_quantizers.188 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.48e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.gate_up_proj_weight_quantizers.189 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.47e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.gate_up_proj_weight_quantizers.190 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.28e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.gate_up_proj_weight_quantizers.191 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.21e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.gate_up_proj_weight_quantizers.192 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.39e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.gate_up_proj_weight_quantizers.193 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.16e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.gate_up_proj_weight_quantizers.194 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.38e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.gate_up_proj_weight_quantizers.195 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.39e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.gate_up_proj_weight_quantizers.196 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.17e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.gate_up_proj_weight_quantizers.197 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.44e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.gate_up_proj_weight_quantizers.198 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.72e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.gate_up_proj_weight_quantizers.199 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.28e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.gate_up_proj_weight_quantizers.200 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.99e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.gate_up_proj_weight_quantizers.201 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.24e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.gate_up_proj_weight_quantizers.202 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.30e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.gate_up_proj_weight_quantizers.203 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.49e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.gate_up_proj_weight_quantizers.204 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.42e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.gate_up_proj_weight_quantizers.205 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.gate_up_proj_weight_quantizers.206 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.gate_up_proj_weight_quantizers.207 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.15e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.gate_up_proj_weight_quantizers.208 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.79e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.gate_up_proj_weight_quantizers.209 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.47e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.gate_up_proj_weight_quantizers.210 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.60e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.gate_up_proj_weight_quantizers.211 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.45e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.gate_up_proj_weight_quantizers.212 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.64e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.gate_up_proj_weight_quantizers.213 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.09e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.gate_up_proj_weight_quantizers.214 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.68e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.gate_up_proj_weight_quantizers.215 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.29e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.gate_up_proj_weight_quantizers.216 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.32e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.gate_up_proj_weight_quantizers.217 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.96e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.gate_up_proj_weight_quantizers.218 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.87e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.gate_up_proj_weight_quantizers.219 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.07e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.gate_up_proj_weight_quantizers.220 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.31e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.gate_up_proj_weight_quantizers.221 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.80e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.gate_up_proj_weight_quantizers.222 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.09e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.gate_up_proj_weight_quantizers.223 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.35e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.gate_up_proj_weight_quantizers.224 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.41e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.gate_up_proj_weight_quantizers.225 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.25e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.gate_up_proj_weight_quantizers.226 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.82e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.gate_up_proj_weight_quantizers.227 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.27e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.gate_up_proj_weight_quantizers.228 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.18e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.gate_up_proj_weight_quantizers.229 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.22e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.gate_up_proj_weight_quantizers.230 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.36e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.gate_up_proj_weight_quantizers.231 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.23e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.gate_up_proj_weight_quantizers.232 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.76e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.gate_up_proj_weight_quantizers.233 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.56e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.gate_up_proj_weight_quantizers.234 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.40e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.gate_up_proj_weight_quantizers.235 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.49e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.gate_up_proj_weight_quantizers.236 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.20e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.gate_up_proj_weight_quantizers.237 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.21e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.gate_up_proj_weight_quantizers.238 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.14e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.gate_up_proj_weight_quantizers.239 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.64e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.gate_up_proj_weight_quantizers.240 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.97e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.gate_up_proj_weight_quantizers.241 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.37e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.gate_up_proj_weight_quantizers.242 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.99e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.gate_up_proj_weight_quantizers.243 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.65e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.gate_up_proj_weight_quantizers.244 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.46e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.gate_up_proj_weight_quantizers.245 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.22e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.gate_up_proj_weight_quantizers.246 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.74e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.gate_up_proj_weight_quantizers.247 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.93e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.gate_up_proj_weight_quantizers.248 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.91e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.gate_up_proj_weight_quantizers.249 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.59e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.gate_up_proj_weight_quantizers.250 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.34e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.gate_up_proj_weight_quantizers.251 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.40e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.gate_up_proj_weight_quantizers.252 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.31e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.gate_up_proj_weight_quantizers.253 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.22e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.gate_up_proj_weight_quantizers.254 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.91e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.gate_up_proj_weight_quantizers.255 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.01e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.down_proj_input_quantizer TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.36e+03 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.down_proj_weight_quantizers.0 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.48e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.down_proj_weight_quantizers.1 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.14e+00 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.down_proj_weight_quantizers.2 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=6.05e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.down_proj_weight_quantizers.3 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.84e+00 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.down_proj_weight_quantizers.4 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=4.49e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.down_proj_weight_quantizers.5 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.36e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.down_proj_weight_quantizers.6 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.66e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.down_proj_weight_quantizers.7 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.28e+00 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.down_proj_weight_quantizers.8 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=5.16e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.down_proj_weight_quantizers.9 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.83e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.down_proj_weight_quantizers.10 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.12e+00 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.down_proj_weight_quantizers.11 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e+00 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.down_proj_weight_quantizers.12 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=6.76e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.down_proj_weight_quantizers.13 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.28e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.down_proj_weight_quantizers.14 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.30e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.down_proj_weight_quantizers.15 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.32e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.down_proj_weight_quantizers.16 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=7.38e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.down_proj_weight_quantizers.17 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.60e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.down_proj_weight_quantizers.18 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=7.07e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.down_proj_weight_quantizers.19 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.32e+00 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.down_proj_weight_quantizers.20 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.29e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.down_proj_weight_quantizers.21 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.26e+00 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.down_proj_weight_quantizers.22 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=5.86e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.down_proj_weight_quantizers.23 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.31e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.down_proj_weight_quantizers.24 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.27e+00 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.down_proj_weight_quantizers.25 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.63e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.down_proj_weight_quantizers.26 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.73e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.down_proj_weight_quantizers.27 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.down_proj_weight_quantizers.28 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=7.81e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.down_proj_weight_quantizers.29 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.65e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.down_proj_weight_quantizers.30 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.22e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.down_proj_weight_quantizers.31 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=4.41e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.down_proj_weight_quantizers.32 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=5.08e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.down_proj_weight_quantizers.33 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=5.35e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.down_proj_weight_quantizers.34 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=7.97e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.down_proj_weight_quantizers.35 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=6.13e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.down_proj_weight_quantizers.36 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=4.79e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.down_proj_weight_quantizers.37 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.57e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.down_proj_weight_quantizers.38 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=4.38e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.down_proj_weight_quantizers.39 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.35e+00 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.down_proj_weight_quantizers.40 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.82e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.down_proj_weight_quantizers.41 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.69e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.down_proj_weight_quantizers.42 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.45e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.down_proj_weight_quantizers.43 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.07e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.down_proj_weight_quantizers.44 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.77e+00 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.down_proj_weight_quantizers.45 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.66e+00 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.down_proj_weight_quantizers.46 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=4.06e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.down_proj_weight_quantizers.47 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=5.16e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.down_proj_weight_quantizers.48 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=4.49e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.down_proj_weight_quantizers.49 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.63e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.down_proj_weight_quantizers.50 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.14e+00 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.down_proj_weight_quantizers.51 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=4.55e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.down_proj_weight_quantizers.52 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.79e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.down_proj_weight_quantizers.53 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.69e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.down_proj_weight_quantizers.54 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.16e+00 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.down_proj_weight_quantizers.55 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.10e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.down_proj_weight_quantizers.56 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=6.84e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.down_proj_weight_quantizers.57 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.20e+00 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.down_proj_weight_quantizers.58 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.14e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.down_proj_weight_quantizers.59 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=4.47e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.down_proj_weight_quantizers.60 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.23e+00 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.down_proj_weight_quantizers.61 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.22e+00 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.down_proj_weight_quantizers.62 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=5.66e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.down_proj_weight_quantizers.63 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=5.39e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.down_proj_weight_quantizers.64 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.23e+00 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.down_proj_weight_quantizers.65 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.52e+00 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.down_proj_weight_quantizers.66 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=6.29e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.down_proj_weight_quantizers.67 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.50e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.down_proj_weight_quantizers.68 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=5.51e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.down_proj_weight_quantizers.69 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.71e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.down_proj_weight_quantizers.70 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=4.34e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.down_proj_weight_quantizers.71 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=5.08e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.down_proj_weight_quantizers.72 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.42e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.down_proj_weight_quantizers.73 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.39e+00 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.down_proj_weight_quantizers.74 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.34e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.down_proj_weight_quantizers.75 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=5.55e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.down_proj_weight_quantizers.76 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.24e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.down_proj_weight_quantizers.77 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.64e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.down_proj_weight_quantizers.78 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.17e+00 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.down_proj_weight_quantizers.79 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.93e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.down_proj_weight_quantizers.80 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=6.64e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.down_proj_weight_quantizers.81 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.52e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.down_proj_weight_quantizers.82 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=6.09e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.down_proj_weight_quantizers.83 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.52e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.down_proj_weight_quantizers.84 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=5.08e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.down_proj_weight_quantizers.85 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=6.29e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.down_proj_weight_quantizers.86 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.48e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.down_proj_weight_quantizers.87 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=4.96e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.down_proj_weight_quantizers.88 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.35e+00 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.down_proj_weight_quantizers.89 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=7.66e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.down_proj_weight_quantizers.90 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=7.11e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.down_proj_weight_quantizers.91 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.98e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.down_proj_weight_quantizers.92 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=6.72e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.down_proj_weight_quantizers.93 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.31e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.down_proj_weight_quantizers.94 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=4.43e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.down_proj_weight_quantizers.95 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=4.96e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.down_proj_weight_quantizers.96 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=5.78e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.down_proj_weight_quantizers.97 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=5.47e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.down_proj_weight_quantizers.98 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.14e+00 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.down_proj_weight_quantizers.99 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.19e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.down_proj_weight_quantizers.100 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.19e+00 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.down_proj_weight_quantizers.101 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e+00 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.down_proj_weight_quantizers.102 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.50e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.down_proj_weight_quantizers.103 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.33e+00 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.down_proj_weight_quantizers.104 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.52e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.down_proj_weight_quantizers.105 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.93e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.down_proj_weight_quantizers.106 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.14e+00 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.down_proj_weight_quantizers.107 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.83e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.down_proj_weight_quantizers.108 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=4.61e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.down_proj_weight_quantizers.109 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e+00 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.down_proj_weight_quantizers.110 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=7.93e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.down_proj_weight_quantizers.111 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=7.11e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.down_proj_weight_quantizers.112 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=6.17e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.down_proj_weight_quantizers.113 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=4.84e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.down_proj_weight_quantizers.114 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.48e+00 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.down_proj_weight_quantizers.115 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e+00 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.down_proj_weight_quantizers.116 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.96e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.down_proj_weight_quantizers.117 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=7.85e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.down_proj_weight_quantizers.118 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.38e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.down_proj_weight_quantizers.119 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e+00 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.down_proj_weight_quantizers.120 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.54e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.down_proj_weight_quantizers.121 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.05e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.down_proj_weight_quantizers.122 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.34e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.down_proj_weight_quantizers.123 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.29e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.down_proj_weight_quantizers.124 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=6.41e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.down_proj_weight_quantizers.125 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.25e+00 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.down_proj_weight_quantizers.126 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.01e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.down_proj_weight_quantizers.127 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.87e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.down_proj_weight_quantizers.128 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.69e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.down_proj_weight_quantizers.129 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.98e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.down_proj_weight_quantizers.130 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.59e+00 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.down_proj_weight_quantizers.131 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=5.51e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.down_proj_weight_quantizers.132 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=4.20e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.down_proj_weight_quantizers.133 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.45e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.down_proj_weight_quantizers.134 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.57e+00 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.down_proj_weight_quantizers.135 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.99e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.down_proj_weight_quantizers.136 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.71e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.down_proj_weight_quantizers.137 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=4.00e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.down_proj_weight_quantizers.138 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.10e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.down_proj_weight_quantizers.139 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.35e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.down_proj_weight_quantizers.140 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.91e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.down_proj_weight_quantizers.141 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.61e+00 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.down_proj_weight_quantizers.142 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=6.80e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.down_proj_weight_quantizers.143 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=4.65e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.down_proj_weight_quantizers.144 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.05e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.down_proj_weight_quantizers.145 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=6.48e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.down_proj_weight_quantizers.146 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.14e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.down_proj_weight_quantizers.147 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.72e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.down_proj_weight_quantizers.148 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=5.08e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.down_proj_weight_quantizers.149 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=5.70e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.down_proj_weight_quantizers.150 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.36e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.down_proj_weight_quantizers.151 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.69e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.down_proj_weight_quantizers.152 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.55e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.down_proj_weight_quantizers.153 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=5.94e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.down_proj_weight_quantizers.154 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.32e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.down_proj_weight_quantizers.155 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=5.39e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.down_proj_weight_quantizers.156 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.21e+00 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.down_proj_weight_quantizers.157 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.18e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.down_proj_weight_quantizers.158 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.03e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.down_proj_weight_quantizers.159 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.36e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.down_proj_weight_quantizers.160 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.75e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.down_proj_weight_quantizers.161 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.84e+00 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.down_proj_weight_quantizers.162 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=5.16e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.down_proj_weight_quantizers.163 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=5.59e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.down_proj_weight_quantizers.164 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=5.47e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.down_proj_weight_quantizers.165 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.99e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.down_proj_weight_quantizers.166 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e+00 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.down_proj_weight_quantizers.167 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.29e+00 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.down_proj_weight_quantizers.168 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.59e+00 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.down_proj_weight_quantizers.169 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.33e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.down_proj_weight_quantizers.170 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.38e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.down_proj_weight_quantizers.171 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.76e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.down_proj_weight_quantizers.172 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=5.27e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.down_proj_weight_quantizers.173 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.91e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.down_proj_weight_quantizers.174 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.10e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.down_proj_weight_quantizers.175 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.27e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.down_proj_weight_quantizers.176 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=5.43e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.down_proj_weight_quantizers.177 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.91e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.down_proj_weight_quantizers.178 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.05e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.down_proj_weight_quantizers.179 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=4.43e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.down_proj_weight_quantizers.180 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.91e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.down_proj_weight_quantizers.181 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.30e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.down_proj_weight_quantizers.182 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.70e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.down_proj_weight_quantizers.183 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.48e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.down_proj_weight_quantizers.184 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.91e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.down_proj_weight_quantizers.185 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=7.15e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.down_proj_weight_quantizers.186 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.40e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.down_proj_weight_quantizers.187 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.21e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.down_proj_weight_quantizers.188 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.24e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.down_proj_weight_quantizers.189 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.98e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.down_proj_weight_quantizers.190 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=4.71e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.down_proj_weight_quantizers.191 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=4.82e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.down_proj_weight_quantizers.192 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=5.16e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.down_proj_weight_quantizers.193 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.16e+00 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.down_proj_weight_quantizers.194 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.69e+00 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.down_proj_weight_quantizers.195 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.89e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.down_proj_weight_quantizers.196 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.63e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.down_proj_weight_quantizers.197 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=5.43e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.down_proj_weight_quantizers.198 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=5.59e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.down_proj_weight_quantizers.199 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.81e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.down_proj_weight_quantizers.200 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.47e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.down_proj_weight_quantizers.201 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=5.62e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.down_proj_weight_quantizers.202 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.48e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.down_proj_weight_quantizers.203 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.81e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.down_proj_weight_quantizers.204 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.16e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.down_proj_weight_quantizers.205 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.62e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.down_proj_weight_quantizers.206 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.10e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.down_proj_weight_quantizers.207 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.42e+00 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.down_proj_weight_quantizers.208 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=4.24e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.down_proj_weight_quantizers.209 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.69e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.down_proj_weight_quantizers.210 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.57e+00 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.down_proj_weight_quantizers.211 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=6.99e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.down_proj_weight_quantizers.212 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=6.76e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.down_proj_weight_quantizers.213 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=4.59e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.down_proj_weight_quantizers.214 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=4.22e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.down_proj_weight_quantizers.215 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=4.12e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.down_proj_weight_quantizers.216 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=5.00e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.down_proj_weight_quantizers.217 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.69e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.down_proj_weight_quantizers.218 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=5.08e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.down_proj_weight_quantizers.219 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=7.77e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.down_proj_weight_quantizers.220 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.71e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.down_proj_weight_quantizers.221 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.62e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.down_proj_weight_quantizers.222 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=5.47e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.down_proj_weight_quantizers.223 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=6.91e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.down_proj_weight_quantizers.224 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.09e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.down_proj_weight_quantizers.225 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.49e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.down_proj_weight_quantizers.226 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e+00 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.down_proj_weight_quantizers.227 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.34e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.down_proj_weight_quantizers.228 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.77e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.down_proj_weight_quantizers.229 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.46e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.down_proj_weight_quantizers.230 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=5.27e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.down_proj_weight_quantizers.231 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.61e+00 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.down_proj_weight_quantizers.232 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.51e+00 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.down_proj_weight_quantizers.233 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.71e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.down_proj_weight_quantizers.234 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.79e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.down_proj_weight_quantizers.235 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=6.05e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.down_proj_weight_quantizers.236 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.46e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.down_proj_weight_quantizers.237 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=4.69e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.down_proj_weight_quantizers.238 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=4.02e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.down_proj_weight_quantizers.239 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.37e+00 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.down_proj_weight_quantizers.240 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.04e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.down_proj_weight_quantizers.241 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.54e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.down_proj_weight_quantizers.242 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.23e+00 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.down_proj_weight_quantizers.243 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.26e+00 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.down_proj_weight_quantizers.244 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.13e+00 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.down_proj_weight_quantizers.245 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.36e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.down_proj_weight_quantizers.246 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.49e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.down_proj_weight_quantizers.247 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.01e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.down_proj_weight_quantizers.248 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.40e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.down_proj_weight_quantizers.249 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.02e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.down_proj_weight_quantizers.250 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.42e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.down_proj_weight_quantizers.251 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.06e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.down_proj_weight_quantizers.252 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.73e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.down_proj_weight_quantizers.253 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.93e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.down_proj_weight_quantizers.254 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.11e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.down_proj_weight_quantizers.255 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e+00 calibrator=MaxCalibrator quant) +model.layers.77.mlp.shared_experts.gate_proj.input_quantizer TensorQuantizer(disabled) +model.layers.77.mlp.shared_experts.gate_proj.output_quantizer TensorQuantizer(disabled) +model.layers.77.mlp.shared_experts.gate_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.77.mlp.shared_experts.up_proj.input_quantizer TensorQuantizer(disabled) +model.layers.77.mlp.shared_experts.up_proj.output_quantizer TensorQuantizer(disabled) +model.layers.77.mlp.shared_experts.up_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.77.mlp.shared_experts.down_proj.input_quantizer TensorQuantizer(disabled) +model.layers.77.mlp.shared_experts.down_proj.output_quantizer TensorQuantizer(disabled) +model.layers.77.mlp.shared_experts.down_proj.weight_quantizer TensorQuantizer(disabled) +lm_head.input_quantizer TensorQuantizer(disabled) +lm_head.output_quantizer TensorQuantizer(disabled) +lm_head.weight_quantizer TensorQuantizer(disabled) +40971 TensorQuantizers found in model diff --git a/LICENSE b/LICENSE new file mode 100644 index 0000000000000000000000000000000000000000..214d7b4a340469926f63d60f7b81589f9d9903e4 --- /dev/null +++ b/LICENSE @@ -0,0 +1,27 @@ +Copyright (c) 2026 Z.AI + +Permission is hereby granted, free of charge, to any person or entity (the "Licensee") obtaining a copy of this software — including the model weights, parameters, configuration files, inference and training code, and associated documentation (collectively, the "Software") — to deal in the Software without restriction, including without limitation the rights to use, copy, modify, merge, publish, distribute, sublicense, and/or sell copies of the Software; to run, deploy, fine-tune, or otherwise modify the Software and create derivative works from it; and to permit persons to whom the Software is furnished to do so, subject to the following conditions: + +1. The above copyright notice and this permission notice shall be included in all copies or substantial portions of the Software. The Licensee's use of the Software must comply with applicable laws and regulations. + +2. "Model as a Service" means giving a third party access to language model inference or fine-tuning (e.g., via API) in a manner that allows such third party to exercise meaningful control over the inputs, parameters, or training data. This does not include (a) end-user products with model capabilities solely embedded within specific features or harnesses, or (b) mere relaying of requests to models hosted by others. +If the Licensee or any of its affiliates operates a Model as a Service business, and the aggregate revenue of the Licensee and its affiliates exceeds 10 billion US dollars (or the equivalent in other currencies) in total over any consecutive 12 months, the Licensee must pass Z.AI's security review before using the Software or its derivative works for any commercial purpose. The scope and method of the security review shall be reasonably determined by Z.AI. + +3. THE SOFTWARE AND ANY OUTPUT AND RESULTS THEREFROM ARE PROVIDED ON AN "AS IS" BASIS, WITHOUT WARRANTY OF ANY KIND, EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL Z.AI OR ITS AFFILIATES OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE. + +For any questions regarding this license, please contact glmlicense@z.ai. + +----- + +版权所有 (c) 2026 Z.AI + +特此免费授予任何获得本软件副本的个人或实体（"被许可方"）——包括模型权重、参数、配置文件、推理和训练代码及相关文档（统称"软件"）——不受限制地处理本软件的权利，包括但不限于：使用、复制、修改、合并、发布、分发、再许可和/或销售软件副本；运行、部署、微调或以其他方式修改软件并创建衍生作品；以及允许获得软件的其他人行使上述权利，但须遵守以下条件： + +1. 上述版权声明和本许可声明应包含在软件的所有副本或实质性部分中。被许可方对软件的使用必须符合适用法律法规。 + +2. "模型即服务"指以允许第三方对输入、参数或训练数据行使实质性控制的方式，向第三方提供语言模型推理或微调服务（如通过API）。不包括：(a) 模型能力仅嵌入特定功能或框架中的终端用户产品，或(b) 单纯转发请求至他人托管的模型。 +若被许可方或其关联方运营"模型即服务"业务，且被许可方及关联方在任意连续12个月内累计总收入超过100亿美元（或等值其他货币），则被许可方在使用软件或其衍生作品进行任何商业用途之前，须通过Z.AI的安全审查。安全审查的范围和方式由Z.AI合理确定。 + +3. 软件及其任何输出和结果均按"现状"提供，不附带任何形式的保证，无论是明示还是暗示，包括但不限于适销性、特定用途适用性和不侵权的保证。在任何情况下，Z.AI或其关联方或版权持有人均不对任何索赔、损害或其他责任承担责任，无论该责任是基于合同、侵权或其他方式，因软件或使用软件而产生或与之相关。 + +如对本许可有任何疑问，请联系 glmlicense@z.ai。 diff --git a/README.md b/README.md new file mode 100644 index 0000000000000000000000000000000000000000..f8c12e2ba71082e465d743c55009d831137a0f6f --- /dev/null +++ b/README.md @@ -0,0 +1,89 @@ +--- +pipeline_tag: text-generation +base_model: +- zai-org/GLM-5.3 +base_model_relation: quantized +license: other +license_name: zai-model-license +license_link: https://huggingface.co/zai-org/GLM-5.3/blob/main/LICENSE +library_name: transformers +tags: +- AxionML +- ModelOpt +- GLM-5 +- quantized +- NVFP4 +- nvfp4 +- sglang +- mirror +--- + +# AxionML GLM-5.3-NVFP4 + +> Mirrored by [AxionML](https://huggingface.co/AxionML) for open-source serving and deployment use cases. Part of AxionML's effort to provide ready-to-serve quantized models for the community. +> +> **Quantized by RadixArk.** The weights in this repository are an unmodified copy of [RadixArk/GLM-5.3-NVFP4](https://huggingface.co/RadixArk/GLM-5.3-NVFP4) (revision `6e389189d564d9d3e4ea7fa284fe91f136d8ae2a`). All credit for the quantization belongs to RadixArk. + +This is an NVFP4-quantized version of [zai-org/GLM-5.3](https://huggingface.co/zai-org/GLM-5.3) (753B total parameters, ~40B activated), quantized with [NVIDIA Model Optimizer](https://github.com/NVIDIA/Model-Optimizer). + +**About NVFP4 quantization:** NVFP4 on Blackwell couples a compact E2M1 FP4 codebook with blockwise FP8 (E4M3) scaling over 16-element micro-blocks, so that 4-bit stored values remain numerically useful for neural-network computation. The E2M1 codebook provides a small, nonuniform set of representable magnitudes up to ±6 and relies on saturating behavior rather than IEEE NaN/Inf encodings to maximize usable range per bit. Using an FP8 block scale (rather than power-of-two-only E8M0) enables fractional scales and error-minimizing scale selection. On Blackwell Tensor Cores, native FP4 multipliers exploit E2M1 simplicity while higher-precision FP32 accumulation protects dot-product accuracy. + +> **Ready for commercial and non-commercial use under the Z.AI Model License** (MIT-style, included as `LICENSE`). + +## Model Summary + +| | | +|:---|:---| +| **Architecture** | Sparse MoE with sparse attention (`GlmMoeDsaForCausalLM`), IndexShare indexer | +| **Total Parameters** | 753B | +| **Activated Parameters** | ~40B | +| **Layers / Experts** | 78 layers (3 dense + 75 MoE), 256 routed experts (top-8) + 1 shared, 1 MTP layer | +| **Context Length** | 1,048,576 tokens | +| **Checkpoint Size** | ~465 GB (vs ~1,507 GB BF16) | + +## Evaluation Results + +| Benchmark | Protocol | NVFP4 | +|:---|:---|:---:| +| GSM8K | Full 1,319-example split | 97.42 | +| Terminal-Bench 2.1 | terminus-2 harness | 86.5 | +| DeepSWE | mini-swe-agent harness | 68.1 | +| AIME 2026 | 30 problems × 16, pass@1 | 94.17 (maj@16 100) | + +> Scores reported by RadixArk for this checkpoint. Against the BF16 source under the same protocol, GSM8K matched exactly and AIME 2026 pass@1 was within run-to-run noise. + +## Quantization Details + +- **Quantization format:** NVFP4 W4A4 (group size 16, FP8 E4M3 block scales, static per-tensor activation scales) on the routed experts of all 75 MoE layers — 96.2% of parameters +- **Unchanged (BF16):** sparse attention incl. IndexShare indexer, shared experts, routers, the 3 dense MLP layers, norms, embeddings, `lm_head`, all MTP tensors +- **Calibration dataset:** 1,024 samples at length 512 from `cnn_dailymail` + `Nemotron-Post-Training-Dataset-v2`, max calibration +- **Tool:** [NVIDIA Model Optimizer](https://github.com/NVIDIA/Model-Optimizer) v0.47.0.dev91 + +## Usage + +### Deploy with SGLang + +```bash +sglang serve \ + --model-path AxionML/GLM-5.3-NVFP4 \ + --tp-size 8 \ + --quantization modelopt_fp4 \ + --reasoning-parser glm45 \ + --tool-call-parser glm47 \ + --speculative-algorithm EAGLE \ + --speculative-num-steps 5 \ + --speculative-eagle-topk 1 \ + --speculative-num-draft-tokens 6 +``` + +Validated upstream on 8x B300. The MTP layer is kept in BF16, so EAGLE speculative decoding works. See the [SGLang GLM-5.3 cookbook](https://cookbook.sglang.io/autoregressive/GLM/GLM-5.3). + +## Limitations + +The base model was trained on data that may contain toxic language and societal biases. The quantized model inherits these limitations. It may generate inaccurate, biased, or offensive content. Please refer to the [original model card](https://huggingface.co/zai-org/GLM-5.3) and the [upstream quantized model card](https://huggingface.co/RadixArk/GLM-5.3-NVFP4) for full details. + +## Credits + +- **Base model:** [zai-org/GLM-5.3](https://huggingface.co/zai-org/GLM-5.3) +- **Quantization:** [RadixArk/GLM-5.3-NVFP4](https://huggingface.co/RadixArk/GLM-5.3-NVFP4) by RadixArk +- **Mirror:** [AxionML](https://huggingface.co/AxionML) diff --git a/chat_template.jinja b/chat_template.jinja new file mode 100644 index 0000000000000000000000000000000000000000..9aff9c44583c869853049724ffe67b1070344f2d --- /dev/null +++ b/chat_template.jinja @@ -0,0 +1,255 @@ +[gMASK] +{%- set effective_reasoning_effort = reasoning_effort if reasoning_effort is defined and reasoning_effort in ['low', 'high'] else 'max' -%} +{%- if effective_reasoning_effort is not none -%}<|system|>Reasoning Effort: {{ effective_reasoning_effort | capitalize }}{%- endif -%} +{%- set clear_thinking = clear_thinking if clear_thinking is defined else false -%} +{%- if tools -%} +{%- macro tool_to_json(tool) -%} + {%- set ns_tool = namespace(first=true) -%} + {{ '{' -}} + {%- for k, v in tool.items() -%} + {%- if k != 'defer_loading' and k != 'strict' -%} + {%- if not ns_tool.first -%}{{- ', ' -}}{%- endif -%} + {%- set ns_tool.first = false -%} + "{{ k }}": {{ v | tojson(ensure_ascii=False) }} + {%- endif -%} + {%- endfor -%} + {{- '}' -}} +{%- endmacro -%} +{%- macro tool_references_to_response(refs) -%} + {{- '\n' -}} + {%- for tr in refs -%} + {%- for tool in tools -%} + {%- if 'function' in tool -%} + {%- set tool = tool['function'] -%} + {%- endif -%} + {%- if tool.name == tr.name -%} + {{- tool_to_json(tool) + '\n' -}} + {%- endif -%} + {%- endfor -%} + {%- endfor -%} + {{- '' -}} +{%- endmacro -%} +<|system|> +# Tools + +You may call one or more functions to assist with the user query. + +You are provided with function signatures within XML tags: + +{% for tool in tools %} +{%- if 'function' in tool -%} + {%- set tool = tool['function'] -%} +{%- endif -%} +{% if tool.defer_loading is not defined or not tool.defer_loading %} +{{ tool_to_json(tool) }} +{% endif %} +{% endfor %} + + +For each function call, output the function name and arguments within the following XML format: +{function-name}{arg-key-1}{arg-value-1}{arg-key-2}{arg-value-2}...{%- endif -%} +{%- macro visible_text(content) -%} + {%- if content is string -%} + {{- content }} + {%- elif content is iterable and content is not mapping -%} + {%- for item in content -%} + {%- if item is mapping and item.type == 'text' -%} + {{- item.text }} + {%- elif item is string -%} + {{- item }} + {%- elif item is mapping and item.type in ['image', 'image_url', 'video', 'video_url', 'audio', 'audio_url', 'input_audio'] -%} + {%- set media_type = item.type | replace('_url', '') | replace('input_', '') -%} + {{- "You are unable to process this " ~ media_type ~ " because you don't have multi-modal input ability. Try different methods." }} + {%- endif -%} + {%- endfor -%} + {%- elif content is not none -%} + {{- content }} + {%- endif -%} +{%- endmacro -%} +{%- macro tool_response(text) -%} +{{- '' + text + '' -}} +{%- endmacro -%} +{%- macro render_tool_response(m) -%} +{%- if m.content is string -%} + {{- tool_response(m.content) -}} +{%- elif m.content and m.content is not mapping and m.content.0.type == "tool_reference" -%} + {{- tool_references_to_response(m.content) -}} +{%- elif is_list_of_outputs(m) -%} + {%- for tr in m.content -%} + {%- if tr.output is iterable and tr.output is not string and tr.output is not mapping and tr.output and tr.output.0.type == "tool_reference" -%} + {{- tool_references_to_response(tr.output) -}} + {%- else -%} + {{- tool_response(visible_text(tr.output)) -}} + {%- endif -%} + {%- endfor -%} +{%- else -%} + {{- tool_response(visible_text(m.content)) -}} +{%- endif -%} +{%- endmacro -%} +{%- macro id_of(obj) -%} + {%- if obj.tool_call_id -%} + {{- obj.tool_call_id -}} + {%- elif obj.id -%} + {{- obj.id -}} + {%- endif -%} +{%- endmacro -%} +{%- macro is_list_of_outputs(m) -%} + {%- if m.content and m.content.0.output is defined -%}1{%- endif -%} +{%- endmacro -%} +{%- macro has_dup_tool_result_id(lo, hi, target) -%} + {%- set ns_cnt = namespace(n=0) -%} + {%- for k in range(lo, hi + 1) -%} + {%- set m = messages[k] -%} + {%- if is_list_of_outputs(m) -%} + {%- for entry in m.content -%} + {%- if id_of(entry) == target -%} + {%- set ns_cnt.n = ns_cnt.n + 1 -%} + {%- endif -%} + {%- endfor -%} + {%- elif id_of(m) == target -%} + {%- set ns_cnt.n = ns_cnt.n + 1 -%} + {%- endif -%} + {%- if ns_cnt.n > 1 -%}{%- break -%}{%- endif -%} + {%- endfor -%} + {%- if ns_cnt.n > 1 -%}1{%- endif -%} +{%- endmacro -%} +{%- macro tc_id_exists(tcs, target) -%} + {%- set ns_f = namespace(found=false) -%} + {%- for tc in tcs -%} + {%- if id_of(tc) == target -%} + {%- set ns_f.found = true -%} + {%- break -%} + {%- endif -%} + {%- endfor -%} + {%- if ns_f.found -%}1{%- endif -%} +{%- endmacro -%} +{%- set ns = namespace(last_user_index=-1) -%} +{%- for m in messages %} + {%- if m.role == 'user' %} + {%- set ns.last_user_index = loop.index0 -%} + {%- endif %} +{%- endfor %} +{%- for m in messages -%} +{%- if m.role == 'user' -%}<|user|>{{ visible_text(m.content) }} +{%- elif m.role == 'assistant' -%} +<|assistant|> +{%- set content = visible_text(m.content) %} +{%- if m.reasoning_content is string %} + {%- set reasoning_content = m.reasoning_content %} +{%- elif '' in content %} + {%- set reasoning_content = content.split('')[0].split('')[-1] %} + {%- set content = content.split('')[-1] %} +{%- endif %} +{%- if (not clear_thinking or loop.index0 > ns.last_user_index) and reasoning_content is defined -%} +{{ '' + reasoning_content + ''}} +{%- else -%} +{{ '' }} +{%- endif -%} +{%- if content.strip() -%} +{{ content.strip() }} +{%- endif -%} +{% if m.tool_calls %} +{% for tc in m.tool_calls %} +{%- if tc.function %} + {%- set tc = tc.function %} +{%- endif %} +{{- '' + tc.name -}} +{% set _args = tc.arguments %}{% for k, v in _args.items() %}{{ k }}{{ v | tojson(ensure_ascii=False) if v is not string else v }}{% endfor %}{% endfor %} +{% endif %} +{%- elif m.role == 'tool' -%} +{%- if loop.first or (messages[loop.index0 - 1].role != "tool") %} + {{- '<|observation|>' -}} + {%- set block_start = loop.index0 -%} + {%- set ns_blk = namespace(end=block_start) -%} + {%- for j in range(block_start, messages|length) -%} + {%- if messages[j].role == 'tool' -%} + {%- set ns_blk.end = j -%} + {%- else -%} + {%- break -%} + {%- endif -%} + {%- endfor -%} + {%- set ns_a = namespace(tool_calls=none) -%} + {%- if block_start > 0 and messages[block_start - 1].role == 'assistant' and messages[block_start - 1].tool_calls -%} + {%- set ns_a.tool_calls = messages[block_start - 1].tool_calls -%} + {%- endif -%} + {%- set ns_chk = namespace(can_sort=true) -%} + {%- if not ns_a.tool_calls -%} + {%- set ns_chk.can_sort = false -%} + {%- else -%} + {%- for k in range(block_start, ns_blk.end + 1) -%} + {%- if not ns_chk.can_sort -%}{%- break -%}{%- endif -%} + {%- set m = messages[k] -%} + {%- if is_list_of_outputs(m) -%} + {%- for entry in m.content -%} + {%- if not ns_chk.can_sort -%}{%- break -%}{%- endif -%} + {%- set eid = id_of(entry) -%} + {%- if not eid -%} + {%- set ns_chk.can_sort = false -%} + {%- elif has_dup_tool_result_id(block_start, ns_blk.end, eid) -%} + {%- set ns_chk.can_sort = false -%} + {%- elif not tc_id_exists(ns_a.tool_calls, eid) -%} + {%- set ns_chk.can_sort = false -%} + {%- endif -%} + {%- endfor -%} + {%- else -%} + {%- set tk_id = id_of(m) -%} + {%- if not tk_id -%} + {%- set ns_chk.can_sort = false -%} + {%- elif has_dup_tool_result_id(block_start, ns_blk.end, tk_id) -%} + {%- set ns_chk.can_sort = false -%} + {%- elif not tc_id_exists(ns_a.tool_calls, tk_id) -%} + {%- set ns_chk.can_sort = false -%} + {%- endif -%} + {%- endif -%} + {%- endfor -%} + {%- for i in range(ns_a.tool_calls | length) -%} + {%- if not ns_chk.can_sort -%}{%- break -%}{%- endif -%} + {%- set tc_id = id_of(ns_a.tool_calls[i]) -%} + {%- if not tc_id -%} + {%- set ns_chk.can_sort = false -%} + {%- endif -%} + {%- for j in range(i + 1, ns_a.tool_calls | length) -%} + {%- if id_of(ns_a.tool_calls[j]) == tc_id -%} + {%- set ns_chk.can_sort = false -%} + {%- break -%} + {%- endif -%} + {%- endfor -%} + {%- endfor -%} + {%- endif -%} + {%- if ns_chk.can_sort -%} + {%- for tc in ns_a.tool_calls -%} + {%- set tc_id = id_of(tc) -%} + {%- for k in range(block_start, ns_blk.end + 1) -%} + {%- set m = messages[k] -%} + {%- if is_list_of_outputs(m) -%} + {%- for entry in m.content -%} + {%- set eid = id_of(entry) -%} + {%- if eid == tc_id -%} + {%- if entry.output is iterable and entry.output is not string and entry.output is not mapping and entry.output and entry.output.0.type == "tool_reference" -%} + {{- tool_references_to_response(entry.output) -}} + {%- else -%} + {{- tool_response(visible_text(entry.output)) -}} + {%- endif -%} + {%- endif -%} + {%- endfor -%} + {%- else -%} + {%- set tk_id = id_of(m) -%} + {%- if tk_id == tc_id -%} + {{- render_tool_response(m) -}} + {%- endif -%} + {%- endif -%} + {%- endfor -%} + {%- endfor -%} + {%- else -%} + {%- for k in range(block_start, ns_blk.end + 1) -%} + {{- render_tool_response(messages[k]) -}} + {%- endfor -%} + {%- endif -%} +{% endif -%} +{%- elif m.role == 'system' -%} +<|system|>{{ visible_text(m.content) }} +{%- endif -%} +{%- endfor -%} +{%- if add_generation_prompt -%} + <|assistant|>{{- '' -}} +{%- endif -%} diff --git a/config.json b/config.json new file mode 100644 index 0000000000000000000000000000000000000000..a041487a1ccf9cf8b05b74a889e6bb924974ec23 --- /dev/null +++ b/config.json @@ -0,0 +1,571 @@ +{ + "architectures": [ + "GlmMoeDsaForCausalLM" + ], + "attention_bias": false, + "attention_dropout": 0.0, + "bos_token_id": 0, + "dtype": "bfloat16", + "eos_token_id": [ + 154820, + 154827, + 154829 + ], + "ep_size": 1, + "first_k_dense_replace": 3, + "head_dim": 64, + "hidden_act": "silu", + "hidden_size": 6144, + "index_head_dim": 128, + "index_n_heads": 32, + "index_share_for_mtp_iteration": true, + "index_skip_topk_offset": 3, + "index_topk": 2048, + "index_topk_freq": 4, + "index_topk_pattern": null, + "indexer_rope_interleave": true, + "indexer_types": [ + "full", + "full", + "full", + "shared", + "shared", + "shared", + "full", + "shared", + "shared", + "shared", + "full", + "shared", + "shared", + "shared", + "full", + "shared", + "shared", + "shared", + "full", + "shared", + "shared", + "shared", + "full", + "shared", + "shared", + "shared", + "full", + "shared", + "shared", + "shared", + "full", + "shared", + "shared", + "shared", + "full", + "shared", + "shared", + "shared", + "full", + "shared", + "shared", + "shared", + "full", + "shared", + "shared", + "shared", + "full", + "shared", + "shared", + "shared", + "full", + "shared", + "shared", + "shared", + "full", + "shared", + "shared", + "shared", + "full", + "shared", + "shared", + "shared", + "full", + "shared", + "shared", + "shared", + "full", + "shared", + "shared", + "shared", + "full", + "shared", + "shared", + "shared", + "full", + "shared", + "shared", + "shared" + ], + "initializer_range": 0.02, + "intermediate_size": 12288, + "kv_lora_rank": 512, + "layer_types": [ + "deepseek_sparse_attention", + "deepseek_sparse_attention", + "deepseek_sparse_attention", + "deepseek_sparse_attention", + "deepseek_sparse_attention", + "deepseek_sparse_attention", + "deepseek_sparse_attention", + "deepseek_sparse_attention", + "deepseek_sparse_attention", + "deepseek_sparse_attention", + "deepseek_sparse_attention", + "deepseek_sparse_attention", + "deepseek_sparse_attention", + "deepseek_sparse_attention", + "deepseek_sparse_attention", + "deepseek_sparse_attention", + "deepseek_sparse_attention", + "deepseek_sparse_attention", + "deepseek_sparse_attention", + "deepseek_sparse_attention", + "deepseek_sparse_attention", + "deepseek_sparse_attention", + "deepseek_sparse_attention", + "deepseek_sparse_attention", + "deepseek_sparse_attention", + "deepseek_sparse_attention", + "deepseek_sparse_attention", + "deepseek_sparse_attention", + "deepseek_sparse_attention", + "deepseek_sparse_attention", + "deepseek_sparse_attention", + "deepseek_sparse_attention", + "deepseek_sparse_attention", + "deepseek_sparse_attention", + "deepseek_sparse_attention", + "deepseek_sparse_attention", + "deepseek_sparse_attention", + "deepseek_sparse_attention", + "deepseek_sparse_attention", + "deepseek_sparse_attention", + "deepseek_sparse_attention", + "deepseek_sparse_attention", + "deepseek_sparse_attention", + "deepseek_sparse_attention", + "deepseek_sparse_attention", + "deepseek_sparse_attention", + "deepseek_sparse_attention", + "deepseek_sparse_attention", + "deepseek_sparse_attention", + "deepseek_sparse_attention", + "deepseek_sparse_attention", + "deepseek_sparse_attention", + "deepseek_sparse_attention", + "deepseek_sparse_attention", + "deepseek_sparse_attention", + "deepseek_sparse_attention", + "deepseek_sparse_attention", + "deepseek_sparse_attention", + "deepseek_sparse_attention", + "deepseek_sparse_attention", + "deepseek_sparse_attention", + "deepseek_sparse_attention", + "deepseek_sparse_attention", + "deepseek_sparse_attention", + "deepseek_sparse_attention", + "deepseek_sparse_attention", + "deepseek_sparse_attention", + "deepseek_sparse_attention", + "deepseek_sparse_attention", + "deepseek_sparse_attention", + "deepseek_sparse_attention", + "deepseek_sparse_attention", + "deepseek_sparse_attention", + "deepseek_sparse_attention", + "deepseek_sparse_attention", + "deepseek_sparse_attention", + "deepseek_sparse_attention", + "deepseek_sparse_attention" + ], + "max_position_embeddings": 1048576, + "mlp_bias": false, + "mlp_layer_types": [ + "dense", + "dense", + "dense", + "sparse", + "sparse", + "sparse", + "sparse", + "sparse", + "sparse", + "sparse", + "sparse", + "sparse", + "sparse", + "sparse", + "sparse", + "sparse", + "sparse", + "sparse", + "sparse", + "sparse", + "sparse", + "sparse", + "sparse", + "sparse", + "sparse", + "sparse", + "sparse", + "sparse", + "sparse", + "sparse", + "sparse", + "sparse", + "sparse", + "sparse", + "sparse", + "sparse", + "sparse", + "sparse", + "sparse", + "sparse", + "sparse", + "sparse", + "sparse", + "sparse", + "sparse", + "sparse", + "sparse", + "sparse", + "sparse", + "sparse", + "sparse", + "sparse", + "sparse", + "sparse", + "sparse", + "sparse", + "sparse", + "sparse", + "sparse", + "sparse", + "sparse", + "sparse", + "sparse", + "sparse", + "sparse", + "sparse", + "sparse", + "sparse", + "sparse", + "sparse", + "sparse", + "sparse", + "sparse", + "sparse", + "sparse", + "sparse", + "sparse", + "sparse" + ], + "model_type": "glm_moe_dsa", + "moe_intermediate_size": 2048, + "moe_layer_freq": 1, + "moe_router_dtype": "float32", + "n_group": 1, + "n_routed_experts": 256, + "n_shared_experts": 1, + "norm_topk_prob": true, + "num_attention_heads": 64, + "num_experts_per_tok": 8, + "num_hidden_layers": 78, + "num_key_value_heads": 64, + "num_nextn_predict_layers": 1, + "pad_token_id": 154820, + "pretraining_tp": 1, + "q_lora_rank": 2048, + "qk_head_dim": 256, + "qk_nope_head_dim": 192, + "qk_rope_head_dim": 64, + "rms_norm_eps": 1e-05, + "rope_interleave": true, + "rope_parameters": { + "rope_theta": 8000000, + "rope_type": "default" + }, + "routed_scaling_factor": 2.5, + "scoring_func": "sigmoid", + "tie_word_embeddings": false, + "topk_group": 1, + "topk_method": "noaux_tc", + "transformers_version": "5.14.1", + "use_cache": true, + "v_head_dim": 256, + "vocab_size": 154880, + "quantization_config": { + "config_groups": { + "group_0": { + "input_activations": { + "dynamic": false, + "num_bits": 4, + "type": "float", + "group_size": 16 + }, + "weights": { + "dynamic": false, + "num_bits": 4, + "type": "float", + "group_size": 16 + }, + "targets": [ + "Linear" + ] + } + }, + "ignore": [ + "lm_head", + "model.embed_tokens", + "model.layers.0*", + "model.layers.1.*", + "model.layers.10.mlp.gate", + "model.layers.10.mlp.shared_experts*", + "model.layers.10.self_attn*", + "model.layers.11.mlp.gate", + "model.layers.11.mlp.shared_experts*", + "model.layers.11.self_attn*", + "model.layers.12.mlp.gate", + "model.layers.12.mlp.shared_experts*", + "model.layers.12.self_attn*", + "model.layers.13.mlp.gate", + "model.layers.13.mlp.shared_experts*", + "model.layers.13.self_attn*", + "model.layers.14.mlp.gate", + "model.layers.14.mlp.shared_experts*", + "model.layers.14.self_attn*", + "model.layers.15.mlp.gate", + "model.layers.15.mlp.shared_experts*", + "model.layers.15.self_attn*", + "model.layers.16.mlp.gate", + "model.layers.16.mlp.shared_experts*", + "model.layers.16.self_attn*", + "model.layers.17.mlp.gate", + "model.layers.17.mlp.shared_experts*", + "model.layers.17.self_attn*", + "model.layers.18.mlp.gate", + "model.layers.18.mlp.shared_experts*", + "model.layers.18.self_attn*", + "model.layers.19.mlp.gate", + "model.layers.19.mlp.shared_experts*", + "model.layers.19.self_attn*", + "model.layers.2.*", + "model.layers.20.mlp.gate", + "model.layers.20.mlp.shared_experts*", + "model.layers.20.self_attn*", + "model.layers.21.mlp.gate", + "model.layers.21.mlp.shared_experts*", + "model.layers.21.self_attn*", + "model.layers.22.mlp.gate", + "model.layers.22.mlp.shared_experts*", + "model.layers.22.self_attn*", + "model.layers.23.mlp.gate", + "model.layers.23.mlp.shared_experts*", + "model.layers.23.self_attn*", + "model.layers.24.mlp.gate", + "model.layers.24.mlp.shared_experts*", + "model.layers.24.self_attn*", + "model.layers.25.mlp.gate", + "model.layers.25.mlp.shared_experts*", + "model.layers.25.self_attn*", + "model.layers.26.mlp.gate", + "model.layers.26.mlp.shared_experts*", + "model.layers.26.self_attn*", + "model.layers.27.mlp.gate", + "model.layers.27.mlp.shared_experts*", + "model.layers.27.self_attn*", + "model.layers.28.mlp.gate", + "model.layers.28.mlp.shared_experts*", + "model.layers.28.self_attn*", + "model.layers.29.mlp.gate", + "model.layers.29.mlp.shared_experts*", + "model.layers.29.self_attn*", + "model.layers.3.mlp.gate", + "model.layers.3.mlp.shared_experts*", + "model.layers.3.self_attn*", + "model.layers.30.mlp.gate", + "model.layers.30.mlp.shared_experts*", + "model.layers.30.self_attn*", + "model.layers.31.mlp.gate", + "model.layers.31.mlp.shared_experts*", + "model.layers.31.self_attn*", + "model.layers.32.mlp.gate", + "model.layers.32.mlp.shared_experts*", + "model.layers.32.self_attn*", + "model.layers.33.mlp.gate", + "model.layers.33.mlp.shared_experts*", + "model.layers.33.self_attn*", + "model.layers.34.mlp.gate", + "model.layers.34.mlp.shared_experts*", + "model.layers.34.self_attn*", + "model.layers.35.mlp.gate", + "model.layers.35.mlp.shared_experts*", + "model.layers.35.self_attn*", + "model.layers.36.mlp.gate", + "model.layers.36.mlp.shared_experts*", + "model.layers.36.self_attn*", + "model.layers.37.mlp.gate", + "model.layers.37.mlp.shared_experts*", + "model.layers.37.self_attn*", + "model.layers.38.mlp.gate", + "model.layers.38.mlp.shared_experts*", + "model.layers.38.self_attn*", + "model.layers.39.mlp.gate", + "model.layers.39.mlp.shared_experts*", + "model.layers.39.self_attn*", + "model.layers.4.mlp.gate", + "model.layers.4.mlp.shared_experts*", + "model.layers.4.self_attn*", + "model.layers.40.mlp.gate", + "model.layers.40.mlp.shared_experts*", + "model.layers.40.self_attn*", + "model.layers.41.mlp.gate", + "model.layers.41.mlp.shared_experts*", + "model.layers.41.self_attn*", + "model.layers.42.mlp.gate", + "model.layers.42.mlp.shared_experts*", + "model.layers.42.self_attn*", + "model.layers.43.mlp.gate", + "model.layers.43.mlp.shared_experts*", + "model.layers.43.self_attn*", + "model.layers.44.mlp.gate", + "model.layers.44.mlp.shared_experts*", + "model.layers.44.self_attn*", + "model.layers.45.mlp.gate", + "model.layers.45.mlp.shared_experts*", + "model.layers.45.self_attn*", + "model.layers.46.mlp.gate", + "model.layers.46.mlp.shared_experts*", + "model.layers.46.self_attn*", + "model.layers.47.mlp.gate", + "model.layers.47.mlp.shared_experts*", + "model.layers.47.self_attn*", + "model.layers.48.mlp.gate", + "model.layers.48.mlp.shared_experts*", + "model.layers.48.self_attn*", + "model.layers.49.mlp.gate", + "model.layers.49.mlp.shared_experts*", + "model.layers.49.self_attn*", + "model.layers.5.mlp.gate", + "model.layers.5.mlp.shared_experts*", + "model.layers.5.self_attn*", + "model.layers.50.mlp.gate", + "model.layers.50.mlp.shared_experts*", + "model.layers.50.self_attn*", + "model.layers.51.mlp.gate", + "model.layers.51.mlp.shared_experts*", + "model.layers.51.self_attn*", + "model.layers.52.mlp.gate", + "model.layers.52.mlp.shared_experts*", + "model.layers.52.self_attn*", + "model.layers.53.mlp.gate", + "model.layers.53.mlp.shared_experts*", + "model.layers.53.self_attn*", + "model.layers.54.mlp.gate", + "model.layers.54.mlp.shared_experts*", + "model.layers.54.self_attn*", + "model.layers.55.mlp.gate", + "model.layers.55.mlp.shared_experts*", + "model.layers.55.self_attn*", + "model.layers.56.mlp.gate", + "model.layers.56.mlp.shared_experts*", + "model.layers.56.self_attn*", + "model.layers.57.mlp.gate", + "model.layers.57.mlp.shared_experts*", + "model.layers.57.self_attn*", + "model.layers.58.mlp.gate", + "model.layers.58.mlp.shared_experts*", + "model.layers.58.self_attn*", + "model.layers.59.mlp.gate", + "model.layers.59.mlp.shared_experts*", + "model.layers.59.self_attn*", + "model.layers.6.mlp.gate", + "model.layers.6.mlp.shared_experts*", + "model.layers.6.self_attn*", + "model.layers.60.mlp.gate", + "model.layers.60.mlp.shared_experts*", + "model.layers.60.self_attn*", + "model.layers.61.mlp.gate", + "model.layers.61.mlp.shared_experts*", + "model.layers.61.self_attn*", + "model.layers.62.mlp.gate", + "model.layers.62.mlp.shared_experts*", + "model.layers.62.self_attn*", + "model.layers.63.mlp.gate", + "model.layers.63.mlp.shared_experts*", + "model.layers.63.self_attn*", + "model.layers.64.mlp.gate", + "model.layers.64.mlp.shared_experts*", + "model.layers.64.self_attn*", + "model.layers.65.mlp.gate", + "model.layers.65.mlp.shared_experts*", + "model.layers.65.self_attn*", + "model.layers.66.mlp.gate", + "model.layers.66.mlp.shared_experts*", + "model.layers.66.self_attn*", + "model.layers.67.mlp.gate", + "model.layers.67.mlp.shared_experts*", + "model.layers.67.self_attn*", + "model.layers.68.mlp.gate", + "model.layers.68.mlp.shared_experts*", + "model.layers.68.self_attn*", + "model.layers.69.mlp.gate", + "model.layers.69.mlp.shared_experts*", + "model.layers.69.self_attn*", + "model.layers.7.mlp.gate", + "model.layers.7.mlp.shared_experts*", + "model.layers.7.self_attn*", + "model.layers.70.mlp.gate", + "model.layers.70.mlp.shared_experts*", + "model.layers.70.self_attn*", + "model.layers.71.mlp.gate", + "model.layers.71.mlp.shared_experts*", + "model.layers.71.self_attn*", + "model.layers.72.mlp.gate", + "model.layers.72.mlp.shared_experts*", + "model.layers.72.self_attn*", + "model.layers.73.mlp.gate", + "model.layers.73.mlp.shared_experts*", + "model.layers.73.self_attn*", + "model.layers.74.mlp.gate", + "model.layers.74.mlp.shared_experts*", + "model.layers.74.self_attn*", + "model.layers.75.mlp.gate", + "model.layers.75.mlp.shared_experts*", + "model.layers.75.self_attn*", + "model.layers.76.mlp.gate", + "model.layers.76.mlp.shared_experts*", + "model.layers.76.self_attn*", + "model.layers.77.mlp.gate", + "model.layers.77.mlp.shared_experts*", + "model.layers.77.self_attn*", + "model.layers.8.mlp.gate", + "model.layers.8.mlp.shared_experts*", + "model.layers.8.self_attn*", + "model.layers.9.mlp.gate", + "model.layers.9.mlp.shared_experts*", + "model.layers.9.self_attn*", + "model.layers.78*" + ], + "quant_algo": "NVFP4", + "kv_cache_scheme": { + "dynamic": false, + "num_bits": 8, + "type": "float" + }, + "producer": { + "name": "modelopt", + "version": "0.47.0.dev91+g7ff81dd79" + }, + "quant_method": "modelopt" + } +} \ No newline at end of file diff --git a/generation_config.json b/generation_config.json new file mode 100644 index 0000000000000000000000000000000000000000..3c938869027ace73ecc9e9018a9705269e92b6cb --- /dev/null +++ b/generation_config.json @@ -0,0 +1,13 @@ +{ + "_from_model_config": true, + "do_sample": true, + "eos_token_id": [ + 154820, + 154827, + 154829 + ], + "pad_token_id": 154820, + "temperature": 1.0, + "top_p": 0.95, + "transformers_version": "5.14.1" +} diff --git a/hf_quant_config.json b/hf_quant_config.json new file mode 100644 index 0000000000000000000000000000000000000000..535ddee77e747b19212fd3ec437a5d8bc8fd8be7 --- /dev/null +++ b/hf_quant_config.json @@ -0,0 +1,244 @@ +{ + "producer": { + "name": "modelopt", + "version": "0.47.0.dev91+g7ff81dd79" + }, + "quantization": { + "quant_algo": "NVFP4", + "kv_cache_quant_algo": "FP8", + "group_size": 16, + "exclude_modules": [ + "lm_head", + "model.embed_tokens", + "model.layers.0*", + "model.layers.1.*", + "model.layers.10.mlp.gate", + "model.layers.10.mlp.shared_experts*", + "model.layers.10.self_attn*", + "model.layers.11.mlp.gate", + "model.layers.11.mlp.shared_experts*", + "model.layers.11.self_attn*", + "model.layers.12.mlp.gate", + "model.layers.12.mlp.shared_experts*", + "model.layers.12.self_attn*", + "model.layers.13.mlp.gate", + "model.layers.13.mlp.shared_experts*", + "model.layers.13.self_attn*", + "model.layers.14.mlp.gate", + "model.layers.14.mlp.shared_experts*", + "model.layers.14.self_attn*", + "model.layers.15.mlp.gate", + "model.layers.15.mlp.shared_experts*", + "model.layers.15.self_attn*", + "model.layers.16.mlp.gate", + "model.layers.16.mlp.shared_experts*", + "model.layers.16.self_attn*", + "model.layers.17.mlp.gate", + "model.layers.17.mlp.shared_experts*", + "model.layers.17.self_attn*", + "model.layers.18.mlp.gate", + "model.layers.18.mlp.shared_experts*", + "model.layers.18.self_attn*", + "model.layers.19.mlp.gate", + "model.layers.19.mlp.shared_experts*", + "model.layers.19.self_attn*", + "model.layers.2.*", + "model.layers.20.mlp.gate", + "model.layers.20.mlp.shared_experts*", + "model.layers.20.self_attn*", + "model.layers.21.mlp.gate", + "model.layers.21.mlp.shared_experts*", + "model.layers.21.self_attn*", + "model.layers.22.mlp.gate", + "model.layers.22.mlp.shared_experts*", + "model.layers.22.self_attn*", + "model.layers.23.mlp.gate", + "model.layers.23.mlp.shared_experts*", + "model.layers.23.self_attn*", + "model.layers.24.mlp.gate", + "model.layers.24.mlp.shared_experts*", + "model.layers.24.self_attn*", + "model.layers.25.mlp.gate", + "model.layers.25.mlp.shared_experts*", + "model.layers.25.self_attn*", + "model.layers.26.mlp.gate", + "model.layers.26.mlp.shared_experts*", + "model.layers.26.self_attn*", + "model.layers.27.mlp.gate", + "model.layers.27.mlp.shared_experts*", + "model.layers.27.self_attn*", + "model.layers.28.mlp.gate", + "model.layers.28.mlp.shared_experts*", + "model.layers.28.self_attn*", + "model.layers.29.mlp.gate", + "model.layers.29.mlp.shared_experts*", + "model.layers.29.self_attn*", + "model.layers.3.mlp.gate", + "model.layers.3.mlp.shared_experts*", + "model.layers.3.self_attn*", + "model.layers.30.mlp.gate", + "model.layers.30.mlp.shared_experts*", + "model.layers.30.self_attn*", + "model.layers.31.mlp.gate", + "model.layers.31.mlp.shared_experts*", + "model.layers.31.self_attn*", + "model.layers.32.mlp.gate", + "model.layers.32.mlp.shared_experts*", + "model.layers.32.self_attn*", + "model.layers.33.mlp.gate", + "model.layers.33.mlp.shared_experts*", + "model.layers.33.self_attn*", + "model.layers.34.mlp.gate", + "model.layers.34.mlp.shared_experts*", + "model.layers.34.self_attn*", + "model.layers.35.mlp.gate", + "model.layers.35.mlp.shared_experts*", + "model.layers.35.self_attn*", + "model.layers.36.mlp.gate", + "model.layers.36.mlp.shared_experts*", + "model.layers.36.self_attn*", + "model.layers.37.mlp.gate", + "model.layers.37.mlp.shared_experts*", + "model.layers.37.self_attn*", + "model.layers.38.mlp.gate", + "model.layers.38.mlp.shared_experts*", + "model.layers.38.self_attn*", + "model.layers.39.mlp.gate", + "model.layers.39.mlp.shared_experts*", + "model.layers.39.self_attn*", + "model.layers.4.mlp.gate", + "model.layers.4.mlp.shared_experts*", + "model.layers.4.self_attn*", + "model.layers.40.mlp.gate", + "model.layers.40.mlp.shared_experts*", + "model.layers.40.self_attn*", + "model.layers.41.mlp.gate", + "model.layers.41.mlp.shared_experts*", + "model.layers.41.self_attn*", + "model.layers.42.mlp.gate", + "model.layers.42.mlp.shared_experts*", + "model.layers.42.self_attn*", + "model.layers.43.mlp.gate", + "model.layers.43.mlp.shared_experts*", + "model.layers.43.self_attn*", + "model.layers.44.mlp.gate", + "model.layers.44.mlp.shared_experts*", + "model.layers.44.self_attn*", + "model.layers.45.mlp.gate", + "model.layers.45.mlp.shared_experts*", + "model.layers.45.self_attn*", + "model.layers.46.mlp.gate", + "model.layers.46.mlp.shared_experts*", + "model.layers.46.self_attn*", + "model.layers.47.mlp.gate", + "model.layers.47.mlp.shared_experts*", + "model.layers.47.self_attn*", + "model.layers.48.mlp.gate", + "model.layers.48.mlp.shared_experts*", + "model.layers.48.self_attn*", + "model.layers.49.mlp.gate", + "model.layers.49.mlp.shared_experts*", + "model.layers.49.self_attn*", + "model.layers.5.mlp.gate", + "model.layers.5.mlp.shared_experts*", + "model.layers.5.self_attn*", + "model.layers.50.mlp.gate", + "model.layers.50.mlp.shared_experts*", + "model.layers.50.self_attn*", + "model.layers.51.mlp.gate", + "model.layers.51.mlp.shared_experts*", + "model.layers.51.self_attn*", + "model.layers.52.mlp.gate", + "model.layers.52.mlp.shared_experts*", + "model.layers.52.self_attn*", + "model.layers.53.mlp.gate", + "model.layers.53.mlp.shared_experts*", + "model.layers.53.self_attn*", + "model.layers.54.mlp.gate", + "model.layers.54.mlp.shared_experts*", + "model.layers.54.self_attn*", + "model.layers.55.mlp.gate", + "model.layers.55.mlp.shared_experts*", + "model.layers.55.self_attn*", + "model.layers.56.mlp.gate", + "model.layers.56.mlp.shared_experts*", + "model.layers.56.self_attn*", + "model.layers.57.mlp.gate", + "model.layers.57.mlp.shared_experts*", + "model.layers.57.self_attn*", + "model.layers.58.mlp.gate", + "model.layers.58.mlp.shared_experts*", + "model.layers.58.self_attn*", + "model.layers.59.mlp.gate", + "model.layers.59.mlp.shared_experts*", + "model.layers.59.self_attn*", + "model.layers.6.mlp.gate", + "model.layers.6.mlp.shared_experts*", + "model.layers.6.self_attn*", + "model.layers.60.mlp.gate", + "model.layers.60.mlp.shared_experts*", + "model.layers.60.self_attn*", + "model.layers.61.mlp.gate", + "model.layers.61.mlp.shared_experts*", + "model.layers.61.self_attn*", + "model.layers.62.mlp.gate", + "model.layers.62.mlp.shared_experts*", + "model.layers.62.self_attn*", + "model.layers.63.mlp.gate", + "model.layers.63.mlp.shared_experts*", + "model.layers.63.self_attn*", + "model.layers.64.mlp.gate", + "model.layers.64.mlp.shared_experts*", + "model.layers.64.self_attn*", + "model.layers.65.mlp.gate", + "model.layers.65.mlp.shared_experts*", + "model.layers.65.self_attn*", + "model.layers.66.mlp.gate", + "model.layers.66.mlp.shared_experts*", + "model.layers.66.self_attn*", + "model.layers.67.mlp.gate", + "model.layers.67.mlp.shared_experts*", + "model.layers.67.self_attn*", + "model.layers.68.mlp.gate", + "model.layers.68.mlp.shared_experts*", + "model.layers.68.self_attn*", + "model.layers.69.mlp.gate", + "model.layers.69.mlp.shared_experts*", + "model.layers.69.self_attn*", + "model.layers.7.mlp.gate", + "model.layers.7.mlp.shared_experts*", + "model.layers.7.self_attn*", + "model.layers.70.mlp.gate", + "model.layers.70.mlp.shared_experts*", + "model.layers.70.self_attn*", + "model.layers.71.mlp.gate", + "model.layers.71.mlp.shared_experts*", + "model.layers.71.self_attn*", + "model.layers.72.mlp.gate", + "model.layers.72.mlp.shared_experts*", + "model.layers.72.self_attn*", + "model.layers.73.mlp.gate", + "model.layers.73.mlp.shared_experts*", + "model.layers.73.self_attn*", + "model.layers.74.mlp.gate", + "model.layers.74.mlp.shared_experts*", + "model.layers.74.self_attn*", + "model.layers.75.mlp.gate", + "model.layers.75.mlp.shared_experts*", + "model.layers.75.self_attn*", + "model.layers.76.mlp.gate", + "model.layers.76.mlp.shared_experts*", + "model.layers.76.self_attn*", + "model.layers.77.mlp.gate", + "model.layers.77.mlp.shared_experts*", + "model.layers.77.self_attn*", + "model.layers.8.mlp.gate", + "model.layers.8.mlp.shared_experts*", + "model.layers.8.self_attn*", + "model.layers.9.mlp.gate", + "model.layers.9.mlp.shared_experts*", + "model.layers.9.self_attn*", + "model.layers.78*" + ] + } +} \ No newline at end of file diff --git a/model-00001-of-00047.safetensors b/model-00001-of-00047.safetensors new file mode 100644 index 0000000000000000000000000000000000000000..95a46170b9b71140d5b486319b0f0ffe13ba6a95 --- /dev/null +++ b/model-00001-of-00047.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:54cfda6ee8ef73cde3d037ef6773159cebff950e782aff40650e543373b0fdd9 +size 9996935800 diff --git a/model-00002-of-00047.safetensors b/model-00002-of-00047.safetensors new file mode 100644 index 0000000000000000000000000000000000000000..0616ffb12e8264c2f98c25edc7cb93d6122737a6 --- /dev/null +++ b/model-00002-of-00047.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:aadf755bbcb4bf3558541bd2570f544717cc104b8f043871aa9cda02f4cbb20f +size 9998076664 diff --git a/model-00003-of-00047.safetensors b/model-00003-of-00047.safetensors new file mode 100644 index 0000000000000000000000000000000000000000..364fbec6a92c0c7aabe4dae3282fa28dbc6efd45 --- /dev/null +++ b/model-00003-of-00047.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:f9f081019c53684056c91ec16cf260ac282a00944e71f40ea2ee0a9807f41649 +size 9995585544 diff --git a/model-00004-of-00047.safetensors b/model-00004-of-00047.safetensors new file mode 100644 index 0000000000000000000000000000000000000000..652e6e9bd7e7ffadf45e09c45e82dc68059f6c3c --- /dev/null +++ b/model-00004-of-00047.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:a753ce45323b95ec75a79c00b2780afd234a5e7e52385fb5537988395816c081 +size 9998076664 diff --git a/model-00005-of-00047.safetensors b/model-00005-of-00047.safetensors new file mode 100644 index 0000000000000000000000000000000000000000..f41322298d93d602449e98e6de27ccf4521effd4 --- /dev/null +++ b/model-00005-of-00047.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:28b880a9fb87c7a59c5920372e4e0e29511473dad49e9fc566abcf0ce145bddd +size 9997417192 diff --git a/model-00006-of-00047.safetensors b/model-00006-of-00047.safetensors new file mode 100644 index 0000000000000000000000000000000000000000..fa3b969bf00c24b88747c74eb1ac7f7da4e17eb2 --- /dev/null +++ b/model-00006-of-00047.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:5f6dd8913622c91b05d444e7870c8b2abe6d3dadcb99593b7b29cadc0b8b0f1d +size 9998081432 diff --git a/model-00007-of-00047.safetensors b/model-00007-of-00047.safetensors new file mode 100644 index 0000000000000000000000000000000000000000..e389ec359233a3da8565b274f24145b6c5dec776 --- /dev/null +++ b/model-00007-of-00047.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:3cc52783fc8e8e4ddb0064954e038fe7820b12d3e7d83342f2cc83d7511e252b +size 9995590872 diff --git a/model-00008-of-00047.safetensors b/model-00008-of-00047.safetensors new file mode 100644 index 0000000000000000000000000000000000000000..f09905abbca6ea9d011b893cf81ca59960042963 --- /dev/null +++ b/model-00008-of-00047.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:e7c99000a9a5294554a7c7d08a98939b51f9d1cd39325c61c30ccd708c166404 +size 9999911744 diff --git a/model-00009-of-00047.safetensors b/model-00009-of-00047.safetensors new file mode 100644 index 0000000000000000000000000000000000000000..e542a541203db77865b1cfc6ef76e627c14d2f1f --- /dev/null +++ b/model-00009-of-00047.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:0bd0c1ea52f6fb9e6deddbb59e545d4378fee4e9577f7c75c28f6136ffbde1c0 +size 9998081760 diff --git a/model-00010-of-00047.safetensors b/model-00010-of-00047.safetensors new file mode 100644 index 0000000000000000000000000000000000000000..a3141062c9f0ce660d8b01030f7f5b073bcbc0f5 --- /dev/null +++ b/model-00010-of-00047.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:7cbf4956d37d46b227acc6b332c22299f08523f8658f91bdd22d2c4fc5e5dac8 +size 9995590752 diff --git a/model-00011-of-00047.safetensors b/model-00011-of-00047.safetensors new file mode 100644 index 0000000000000000000000000000000000000000..f0838b24741cfd61dfae977cc94256c461b9bf82 --- /dev/null +++ b/model-00011-of-00047.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:9f9bbae3acc8b5c53e63bbbdc23912e25bbb8b7f31ccbb83eeb104c6acdbab57 +size 9998081968 diff --git a/model-00012-of-00047.safetensors b/model-00012-of-00047.safetensors new file mode 100644 index 0000000000000000000000000000000000000000..ed0b77a131e91f3c787a5d3f4c2b9fcb03f94032 --- /dev/null +++ b/model-00012-of-00047.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:5fcf103c392fc46e9a10cc4e9594a4a2568616c52dac0727d761cae742fc06e0 +size 9997420424 diff --git a/model-00013-of-00047.safetensors b/model-00013-of-00047.safetensors new file mode 100644 index 0000000000000000000000000000000000000000..5d29fec65ce114b700058139a565df52730a0460 --- /dev/null +++ b/model-00013-of-00047.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:edc68efbcd0eee3d029c35d1ba2abf799da1e7b0891b965da3e01bb7e59b83fc +size 9998081432 diff --git a/model-00014-of-00047.safetensors b/model-00014-of-00047.safetensors new file mode 100644 index 0000000000000000000000000000000000000000..d36884ccb0f3d32e304f6d8838ef11e453322bba --- /dev/null +++ b/model-00014-of-00047.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:49e3b9164f48dc84f626f7294754592b37760628ac6b7137fa68e30632159a20 +size 9995590992 diff --git a/model-00015-of-00047.safetensors b/model-00015-of-00047.safetensors new file mode 100644 index 0000000000000000000000000000000000000000..28d528906709bc11d7fae38623fe1428fa766dc6 --- /dev/null +++ b/model-00015-of-00047.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:a61888c260df7ae6d9746042ab8e8888c3c45dd472095e71f3760b0c6b323d35 +size 9999911624 diff --git a/model-00016-of-00047.safetensors b/model-00016-of-00047.safetensors new file mode 100644 index 0000000000000000000000000000000000000000..83194c5cf57f4feaddac338b0fc1fd07c6435c35 --- /dev/null +++ b/model-00016-of-00047.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:3a997506b60e23e8351cdbf98718bfbe89379d7e8bd32e3cf6a280df3450379d +size 9998081640 diff --git a/model-00017-of-00047.safetensors b/model-00017-of-00047.safetensors new file mode 100644 index 0000000000000000000000000000000000000000..5131c818423bd330c0d2bba93df576571784b4db --- /dev/null +++ b/model-00017-of-00047.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:5d6fe8303a9d34aeab085dea5de27d7bf3b3d70629c23a88c28cde08fecb7a44 +size 9995590752 diff --git a/model-00018-of-00047.safetensors b/model-00018-of-00047.safetensors new file mode 100644 index 0000000000000000000000000000000000000000..17ed18401dfa12f69c53811128425583f8a6bed3 --- /dev/null +++ b/model-00018-of-00047.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:27a1e6bede868422409fa7533f3a133d001553681b18351f59ac04a0e7449d40 +size 9959557432 diff --git a/model-00019-of-00047.safetensors b/model-00019-of-00047.safetensors new file mode 100644 index 0000000000000000000000000000000000000000..bab539547a1798b78472ee132b881845cae60c62 --- /dev/null +++ b/model-00019-of-00047.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:dc4be9191e96ceeb2905d6a9d1860d5537e670aa2601e044719d77e6b90fb156 +size 10000552960 diff --git a/model-00020-of-00047.safetensors b/model-00020-of-00047.safetensors new file mode 100644 index 0000000000000000000000000000000000000000..c59ad5ba6f26e413228dd1b69678f8c50abf3be9 --- /dev/null +++ b/model-00020-of-00047.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:08b4839447961ee017afce5eec5137ebe77fe52b5b349215f02c08895c42911e +size 9998081424 diff --git a/model-00021-of-00047.safetensors b/model-00021-of-00047.safetensors new file mode 100644 index 0000000000000000000000000000000000000000..d263d87d3c5275c8094b1a4872e61b89938403f6 --- /dev/null +++ b/model-00021-of-00047.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:6234e3fb04a5def627ddbe7f1ab1c61fdcb17cf92b39376a3182b98f1fd958ea +size 9995591144 diff --git a/model-00022-of-00047.safetensors b/model-00022-of-00047.safetensors new file mode 100644 index 0000000000000000000000000000000000000000..531fd9d5a999612804078fa4fd5f07fb6d886c47 --- /dev/null +++ b/model-00022-of-00047.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:8d461c49b8146fbd294775a08a3328519d1ca6919c65e20cff30eaa550d9956f +size 9999911480 diff --git a/model-00023-of-00047.safetensors b/model-00023-of-00047.safetensors new file mode 100644 index 0000000000000000000000000000000000000000..ea9d9c0fd5f13150c4de0e3aaed7f5662ad61aa4 --- /dev/null +++ b/model-00023-of-00047.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:6d84bc60d844b260f99cac4bde533ebaddd522fa011d628351867184b4418346 +size 9998081504 diff --git a/model-00024-of-00047.safetensors b/model-00024-of-00047.safetensors new file mode 100644 index 0000000000000000000000000000000000000000..afbf920e5b089af527467fcc84dc62915dbb32b8 --- /dev/null +++ b/model-00024-of-00047.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:849657969c725c9a4265482b9c380c529cde1eac23a3f4c6faf0838ac40712f9 +size 9995590752 diff --git a/model-00025-of-00047.safetensors b/model-00025-of-00047.safetensors new file mode 100644 index 0000000000000000000000000000000000000000..825cff5a725ce229756d3f0895b0ed45e3d7fdc9 --- /dev/null +++ b/model-00025-of-00047.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:3d8e2f8777b33add0617a4a22c58aa932c43b13bab4cc08ccb6a68c835e885e9 +size 9999911904 diff --git a/model-00026-of-00047.safetensors b/model-00026-of-00047.safetensors new file mode 100644 index 0000000000000000000000000000000000000000..806552f275e7d481cc0d31ecb325b97943b61ffc --- /dev/null +++ b/model-00026-of-00047.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:a27dd0d81245ca9c4d233d88636fafcbc71ece1f7b6171c9fbfeb63475701af9 +size 9995591304 diff --git a/model-00027-of-00047.safetensors b/model-00027-of-00047.safetensors new file mode 100644 index 0000000000000000000000000000000000000000..cdc882579a527e71632fa7ffda9238b1ba485d01 --- /dev/null +++ b/model-00027-of-00047.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:d6e6bbc295bfa0cc7364ba8918bb1cec49a2fd9ef933ad859d27b3ca646e75fb +size 9998081416 diff --git a/model-00028-of-00047.safetensors b/model-00028-of-00047.safetensors new file mode 100644 index 0000000000000000000000000000000000000000..8f941048363313dec126d72408db1749955010a5 --- /dev/null +++ b/model-00028-of-00047.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:6b099aac1a6a521d156992c7325d1c107cc60e755e781f786f7a91c93532fdb6 +size 9995591208 diff --git a/model-00029-of-00047.safetensors b/model-00029-of-00047.safetensors new file mode 100644 index 0000000000000000000000000000000000000000..1bc9a11f780bc0317344150f0b485d67f6100696 --- /dev/null +++ b/model-00029-of-00047.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:a2325de35c9a76fba4383a7d8f8a40f589efa2f7b690774dc5d0b9a8c504c705 +size 9999911360 diff --git a/model-00030-of-00047.safetensors b/model-00030-of-00047.safetensors new file mode 100644 index 0000000000000000000000000000000000000000..7413e289ae9610815f58244255874b4521b378d9 --- /dev/null +++ b/model-00030-of-00047.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:ff791c6397203263808d750aa75d07c1fa611dccb3a3a27d7755016f078455e1 +size 9998081432 diff --git a/model-00031-of-00047.safetensors b/model-00031-of-00047.safetensors new file mode 100644 index 0000000000000000000000000000000000000000..9e215741198548561bf496c586ac8cd55c8f78fa --- /dev/null +++ b/model-00031-of-00047.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:225582a5b31a0c56a1cfcc3e575ad9e5fbd52e816947e7614b73709531c9b297 +size 9995590840 diff --git a/model-00032-of-00047.safetensors b/model-00032-of-00047.safetensors new file mode 100644 index 0000000000000000000000000000000000000000..4c969bf17597298e46f30915f363b71b9cf397f1 --- /dev/null +++ b/model-00032-of-00047.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:b3c5fa797d08603a1c4bb14d0fe1f556a3a7ef01e71bf8fd074fe4cdffb7b028 +size 9999911784 diff --git a/model-00033-of-00047.safetensors b/model-00033-of-00047.safetensors new file mode 100644 index 0000000000000000000000000000000000000000..f740d2b2113228bddfcaf0491aa20bd36a28d7a9 --- /dev/null +++ b/model-00033-of-00047.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:62c3910877e80c475643ee66e1e5455f168364d9549a30335e88661e25ce33c3 +size 9995591128 diff --git a/model-00034-of-00047.safetensors b/model-00034-of-00047.safetensors new file mode 100644 index 0000000000000000000000000000000000000000..217f2386dd15c95879bc1a59d4c1bcec2a8108ee --- /dev/null +++ b/model-00034-of-00047.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:892b80290f9b525a298f8419980636adc648781da8c32a5dbc679dd19d16e5a1 +size 9998081416 diff --git a/model-00035-of-00047.safetensors b/model-00035-of-00047.safetensors new file mode 100644 index 0000000000000000000000000000000000000000..53f3cb37303e7afae5d6fa42c2c7d74814d334ec --- /dev/null +++ b/model-00035-of-00047.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:188f41d446af253be9e5905fca76762262d5b0da5ded65996953b4e3174c3659 +size 9995591272 diff --git a/model-00036-of-00047.safetensors b/model-00036-of-00047.safetensors new file mode 100644 index 0000000000000000000000000000000000000000..64aacc39fc545546849ac9129a1a17bc6a093dca --- /dev/null +++ b/model-00036-of-00047.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:bf24fce364731f213b7d8ef94582052245aff3edeb9a56c7ecf1b4653ac8eef3 +size 9999911176 diff --git a/model-00037-of-00047.safetensors b/model-00037-of-00047.safetensors new file mode 100644 index 0000000000000000000000000000000000000000..89cd880c232f51aa581e87903bfee0ab00c081ac --- /dev/null +++ b/model-00037-of-00047.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:5629ae8f3ff9e72221301cd0e4d28d01dfdbaa8b186705b1b9a1dacec9730fdc +size 9998081432 diff --git a/model-00038-of-00047.safetensors b/model-00038-of-00047.safetensors new file mode 100644 index 0000000000000000000000000000000000000000..c6291c95d9504d007f3809f5ce1791b1568c799d --- /dev/null +++ b/model-00038-of-00047.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:52cd26324702945141c0e3423f55f4a2b5e37e1dabad7686922a728b351c969f +size 9995590960 diff --git a/model-00039-of-00047.safetensors b/model-00039-of-00047.safetensors new file mode 100644 index 0000000000000000000000000000000000000000..6e3fd4e70de18f8a7316641cc5e041983a620bc1 --- /dev/null +++ b/model-00039-of-00047.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:d4b28d91524ef0efcb4e41f42e3d875979bf6b587bc701dd0018fe1b2fffca25 +size 9999911664 diff --git a/model-00040-of-00047.safetensors b/model-00040-of-00047.safetensors new file mode 100644 index 0000000000000000000000000000000000000000..aeb5455acd0f0df82cba3739848c8a0f418ea2c4 --- /dev/null +++ b/model-00040-of-00047.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:6ca8518f5c6effc532a62339509835f0a67498b14b3fb99d2300d71e43cf17b8 +size 9995591008 diff --git a/model-00041-of-00047.safetensors b/model-00041-of-00047.safetensors new file mode 100644 index 0000000000000000000000000000000000000000..bd529c33534b5083d9d8610c50c61fbfcee11a4b --- /dev/null +++ b/model-00041-of-00047.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:13125b02b3952412ca462d7f847316769001e2c69fd50f73489adfd2c0a3d348 +size 9998081416 diff --git a/model-00042-of-00047.safetensors b/model-00042-of-00047.safetensors new file mode 100644 index 0000000000000000000000000000000000000000..43ba2ca9ee1aee06bee5af126394ac9c155b3d15 --- /dev/null +++ b/model-00042-of-00047.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:bbbeb927147c9b9d54ce8a39eb16869ff21dff00382b3f70859201e056916621 +size 9881695296 diff --git a/model-00043-of-00047.safetensors b/model-00043-of-00047.safetensors new file mode 100644 index 0000000000000000000000000000000000000000..94c4a9cf0b9d9d1cc66d604eb4376a4f28580be7 --- /dev/null +++ b/model-00043-of-00047.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:eb2814f7eb6466319bbf53827f21baa8618a7067ddd000693b68f0bee4624ded +size 10000552944 diff --git a/model-00044-of-00047.safetensors b/model-00044-of-00047.safetensors new file mode 100644 index 0000000000000000000000000000000000000000..81207c213e7dcce6ce79c1e7c2e0252effe9c8e1 --- /dev/null +++ b/model-00044-of-00047.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:5000bbd83f5553e19eda0a5eae762532e5dfbeb998df4d18fc6fab5c2addc2a6 +size 9998081424 diff --git a/model-00045-of-00047.safetensors b/model-00045-of-00047.safetensors new file mode 100644 index 0000000000000000000000000000000000000000..57efbc24646151c8bd07ca51609db69b6368b7bd --- /dev/null +++ b/model-00045-of-00047.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:e16ed8460b75ed61c62add5a4bab0057c37cff53d8d21d3ffa147e0f8f68f485 +size 9980927784 diff --git a/model-00046-of-00047.safetensors b/model-00046-of-00047.safetensors new file mode 100644 index 0000000000000000000000000000000000000000..78801a3493298742994a024b27ceefb6f57df174 --- /dev/null +++ b/model-00046-of-00047.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:f7523821eab72aa188b3e730edc27b15cac0589db11de6c6a220c0740d3b958a +size 9990882136 diff --git a/model-00047-of-00047.safetensors b/model-00047-of-00047.safetensors new file mode 100644 index 0000000000000000000000000000000000000000..1d912eabab78b648fbffbb4fb0fadaa70c0f2d4e --- /dev/null +++ b/model-00047-of-00047.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:bfd0f45a888c1a113407520aa3817f471f50ab267656d1884f48469523d988e2 +size 5108325840 diff --git a/model.safetensors.index.json b/model.safetensors.index.json new file mode 100644 index 0000000000000000000000000000000000000000..0b72f5ba0e99a9b9d31d53191b2eea68dce4e84b --- /dev/null +++ b/model.safetensors.index.json @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:2aa8397b501d9f6a232d153f328feb912f813c389061aac4cf72b04914fa5b74 +size 22194492 diff --git a/tokenizer.json b/tokenizer.json new file mode 100644 index 0000000000000000000000000000000000000000..aba40197a4cdb5607f4ab7a05fb0a4ee8054fd6d --- /dev/null +++ b/tokenizer.json @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:19e773648cb4e65de8660ea6365e10acca112d42a854923df93db4a6f333a82d +size 20217442 diff --git a/tokenizer_config.json b/tokenizer_config.json new file mode 100644 index 0000000000000000000000000000000000000000..0891a18eef6fc3d26dd2e87796ed39c7f91e3a77 --- /dev/null +++ b/tokenizer_config.json @@ -0,0 +1,34 @@ +{ + "backend": "tokenizers", + "clean_up_tokenization_spaces": false, + "do_lower_case": false, + "eos_token": "<|endoftext|>", + "extra_special_tokens": [ + "<|endoftext|>", + "[MASK]", + "[gMASK]", + "[sMASK]", + "", + "", + "<|system|>", + "<|user|>", + "<|assistant|>", + "<|observation|>", + "<|begin_of_image|>", + "<|end_of_image|>", + "<|begin_of_video|>", + "<|end_of_video|>", + "<|begin_of_audio|>", + "<|end_of_audio|>", + "<|begin_of_transcription|>", + "<|end_of_transcription|>" + ], + "is_local": true, + "local_files_only": false, + "model_max_length": 1048576, + "model_specific_special_tokens": {}, + "pad_token": "<|endoftext|>", + "padding_side": "left", + "remove_space": false, + "tokenizer_class": "TokenizersBackend" +}